New energy automobile insurance dynamic risk pricing analysis system

Abstract

The invention relates to the technical field of computers, and discloses a new energy automobile insurance dynamic risk pricing analysis system, which comprises a multi-source acquisition end, a three-dimensional analysis end, a dynamic optimization end, an insurance premium calculation end, an API server end and a risk early warning end, the multi-source acquisition end, the three-dimensional analysis end, the dynamic optimization end, the insurance premium calculation end, the API server end and the risk early warning end are jointly provided with a system monitoring module. By arranging a multi-source acquisition end, when new energy automobile risk analysis is carried out, by obtaining main engine plant battery data, Internet of Vehicles driving behaviors and insurance accident records in real time and establishing a multi-source data fusion channel, real-time interactive verification of data of different sources is ensured, and meanwhile, the state of an automobile is dynamically tracked; the battery health degree attenuation and driving behavior sudden change core risk factors can be detected in real time, the problem of risk mismatching caused by traditional static parameter lag is solved, and the risk identification timeliness and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a dynamic risk pricing analysis system for new energy vehicle insurance. Background Art

[0002] New energy vehicles refer to vehicles that use unconventional automotive fuels as their power source and integrate advanced technologies in vehicle power control and drive to form vehicles with advanced technical principles, new technologies, and new structures. New energy vehicles include pure electric vehicles, extended-range electric vehicles, hybrid vehicles, fuel cell electric vehicles, and hydrogen engine vehicles. With the rapid development of the new energy vehicle industry, and the fact that new energy vehicle owners are tending to be younger, their overall driving experience is insufficient, and driving behavior problems are emerging in an endless stream.

[0003] At present, there are multiple technical bottlenecks in the risk pricing process of new energy vehicle insurance: mainstream actuarial models rely too much on static vehicle parameters and cannot capture the core risks of real-time battery health degradation and dynamic evolution of driving behavior. Although existing technical solutions collect some driving data through on-board equipment, they still face essential limitations: it is impossible to establish a quantitative actuarial mapping mechanism between sudden acceleration and deceleration behavior and premium calculation, resulting in a driving risk premium that deviates from reality; the mileage measurement system does not combine the battery degradation curve for dynamic risk correction, resulting in high-risk vehicles not being reasonably identified; at the same time, because battery status data, intelligent driving system logs, and insurance claims records belong to the independent systems of the OEM, Internet of Vehicles platform, and insurance company, a data collaboration barrier is formed: battery maintenance history data cannot be used in accident probability modeling, which weakens the accuracy of risk prediction; the risk level differences caused by the iteration of the intelligent driving system version are not distinguished, resulting in risk mismatch; dynamic driving behavior factors and static vehicle parameters cannot achieve adaptive weight balance, resulting in a lack of model sensitivity.

[0004] Therefore, a new energy vehicle insurance dynamic risk pricing analysis system is proposed to solve the above problems. Summary of the Invention

[0005] (1) Technical problems solved

[0006] In view of the shortcomings of the existing technology, the present invention provides a new energy vehicle insurance dynamic risk pricing analysis system, which solves the problems raised in the above background technology.

[0007] (2) Technical solution

[0008] To achieve the above objectives, the present invention provides the following technical solutions: a new energy vehicle insurance dynamic risk pricing analysis system, the system comprising a multi-source acquisition terminal, a three-dimensional analysis terminal, a dynamic optimization terminal, a premium calculation terminal, an API server terminal, and a risk warning terminal, wherein the multi-source acquisition terminal, the three-dimensional analysis terminal, the dynamic optimization terminal, the premium calculation terminal, the API server terminal, and the risk warning terminal are jointly provided with a system monitoring module;

[0009] The multi-source acquisition terminal acquires driving data in real time through the vehicle CAN bus interface, retrieves vehicle configuration information from the OEM, establishes an encrypted connection with the insurance company's database, and integrates user feedback data to form a multi-source data set;

[0010] The three-dimensional analysis terminal performs:

[0011] Vehicle condition analysis: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on historical accident counts and average compensation, and identifies intelligent configuration technology risks.

[0012] Driving behavior dimension analysis: Analyze the proportion of sudden acceleration, sudden braking, and sharp turns, calculate the highway usage rate and the proportion of nighttime driving, and generate behavioral style labels based on intelligent driving logs;

[0013] Actuarial dimension modeling: Construct a five-dimensional factor weight system and calculate the initial weights of industry price factor, historical accident factor, battery factor, intelligent factor, and driving behavior factor using the entropy weight method;

[0014] The dynamic optimization end maps user satisfaction feedback to corresponding factor dimensions, iteratively updates factor weight parameters based on a reinforcement learning mechanism, and implements weight lower limit threshold protection for battery factors and risk factors.

[0015] Preferably, the system further comprises:

[0016] The premium calculation terminal maps the standardized factor values ​​of each dimension to a unified scoring range, calculates the recommended premium value based on the dynamic weight, and generates a visual actuarial report including the contribution ratio of each factor;

[0017] The API server pushes the three-dimensional risk assessment report to the insurance pricing system through the OAuth2.0 authentication protocol, and simultaneously supports data call requests from the fleet management platform;

[0018] When the product of the historical risk factor scores is greater than 0.25 and the battery health is less than 70%, the risk warning terminal triggers a high-risk flag to be embedded in the actuarial report, sends a mandatory underwriting instruction to the insurance company, and sets a lower limit for premium output;

[0019] The system monitoring module detects the operating status of each terminal in real time, starts the backup data channel when data is abnormal or service is interrupted, and locates the fault node through the log analysis engine.

[0020] Preferably, the system comprises the following steps:

[0021] S1. Receive vehicle frame number (VIN) input and invoke a multi-source data acquisition module based on the VIN to obtain real-time and historical vehicle data. The data includes OEM vehicle production data, connected vehicle behavior data, insurance company historical claim data, and user feedback data.

[0022] S2. Parallel execution of 3D risk assessment report generation:

[0023] Vehicle Condition Report: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on the number of historical accidents, types, and average compensation, identifies the intelligence level, and predicts technology iteration risks;

[0024] Driving habit analysis: Analyze driving data to calculate the proportion of sudden acceleration, sudden braking, and sharp turns, the rate of highway use, and the proportion of nighttime driving, and generate behavioral style labels based on the intelligent driving system log;

[0025] Actuarial modeling: Construct a five-dimensional factor weighting system: industry price, historical claims, battery, intelligence, and driving behavior factors, and calculate the initial weights using the entropy weight method;

[0026] S3. Perform dynamic weight optimization:

[0027] Map user satisfaction feedback to corresponding factors;

[0028] Iteratively update factor weights based on human feedback reinforcement learning mechanism, and set time decay coefficient to optimize the impact of recent feedback;

[0029] Set lower weight thresholds for battery factors and risk factors;

[0030] S4. Generate premium actuarial recommendations:

[0031] Standardize and map each factor value to a unified scoring interval;

[0032] Calculate recommended premiums based on dynamic weights and standardized scores, and output a visual report containing the contribution values ​​of each factor;

[0033] S5. Push 3D reports to insurance pricing systems and fleet management platforms through API interfaces.

[0034] Preferably, the vehicle condition report generation branch in step S2 specifically includes:

[0035] Calculate real-time health based on battery cycle count and voltage fluctuation data:

[0036]

[0037] Where k is the attenuation coefficient, ΔV is the voltage fluctuation amplitude, V max is the maximum nominal voltage of the battery, and T is the cumulative usage time of the battery;

[0038] Build an intelligent configuration risk matrix to identify sensor configurations with a density below the threshold D min 、OTA update interval exceeds T max Model of car;

[0039] LiDAR models and chip models that are at risk of discontinuation are marked based on the supply chain database.

[0040] Preferably, the behavior style labeling method in the driving habit analysis branch includes:

[0041] Conservative label defined: The intelligent driving system control rate is greater than 70% and the emergency braking ratio is less than 0.1;

[0042] Definition of the aggressive label: the human driving ratio is greater than 60% and the high-speed usage rate is greater than 0.8;

[0043] Define hybrid labels: abnormal exit rate of intelligent driving ∈ [0.2, 0.5] and night driving ratio ∈ [0.3, 0.6]. By acquiring the battery data of the OEM, Internet of Vehicles driving behavior and insurance claim records in real time, and establishing a multi-source data fusion channel, we can ensure real-time interactive verification of data from different sources; at the same time, by dynamically tracking the vehicle status, we can detect the core risk factors of battery health degradation and driving behavior mutation in real time.

[0044] Preferably, the entropy weight method calculation process includes:

[0045] For n samples, the factor data matrix X m×n Normalization processing:

[0046]

[0047] where x ij is the original value of the jth factor of the i-th sample, minx j 、maxx j is the minimum and maximum value;

[0048] Calculate information entropy:

[0049]

[0050] where p ij =x′ ij 、

[0051] Calculate the initial weights:

[0052]

[0053] in is the initial weight of the j-th factor, is the sum of all factor utility values, 1-E j is the information utility value.

[0054] Preferably, the feedback mapping rules of the RLHF engine include:

[0055] When the user marks the premium recommendation as "satisfied", the weight of the factor with a contribution value greater than 10% is increased by ΔW = μ·(1-W t ), where μ is the gain coefficient; when marked as “unsatisfactory”, the factor contribution analysis submodule is triggered to identify the score S i Factor ≤ 0.3 and reduce its weight ΔW = -0.1· W t .

[0056] Preferably, the secondary factors of the five-dimensional factor system include:

[0057] Secondary factor of industry price factor: regional traffic congestion index C I , average annual mileage M avg ;

[0058] Secondary factor of historical accident factor: major accident mark F acc , accident time distribution weight W day / night ;

[0059] Secondary factor of battery factor: battery brand risk coefficient R brand , sensor failure rate λ s ;

[0060] Secondary factor of intelligent factor: ADAS false alarm rate E false , supplier change impact value δ supplier ;

[0061] Secondary factor of driving behavior factor: DMS fatigue driving index I fatigue , extreme weather accident probability P stormo By executing coupled analysis of vehicle condition, driving behavior, and actuarial dimensions in parallel, and constructing a five-dimensional factor weight system based on the entropy weight method; in response to the current situation of multi-source data fragmentation, it realizes real-time correlation between battery maintenance records and accident probability, and analyzes the risk differences of different intelligent driving systems in a graded manner, so that the system can eliminate the data island effect, and automatically correct cross-system data conflicts through the weight optimization mechanism.

[0062] Preferably, the regional traffic congestion index CI The calculation method is:

[0063]

[0064] Where V real is the real-time vehicle speed, I rush is the peak period coefficient, T is the total number of periods in the statistical period, V max The speed limit of the road.

[0065] Preferably, the standardized scoring system in step S4 is defined as:

[0066] Battery Health Rating:

[0067] S batt =SOH / 100

[0068] Driving behavior score:

[0069]

[0070] Intelligent scoring:

[0071]

[0072] Among them S batt SOH is the battery health score, N accel For rapid acceleration, N brake For emergency braking, N turn is the number of sharp turns, N total is the total number of driving behaviors, L level is the intelligent driving level mapping value, Δt update T is the number of days of OTA delay, std This is the standard upgrade cycle;

[0073] High-risk user warning steps:

[0074] When the condition W is met acc ·S acc ≥0.25, S batt When the value is ≤0.7, an early warning signal is generated and at least one of the following actions is triggered:

[0075] Mark high-risk warning signs in actuarial reports;

[0076] Push mandatory underwriting review requests to the insurance company system;

[0077] Limit the lower limit of the premium output value of the API interface to 120% of the base premium;

[0078] Where W acc is the weight of historical risk factors, S accTo score accidents, the battery degradation slope is converted into a dynamic actuarial factor, the contribution of the lidar false detection rate to the accident rate is quantified, and a risk decoupling model for human-machine co-driving is established; when the probability of thermal runaway exceeds the standard, OTA failure, or aggressive driving behavior is identified, the high-risk premium strategy and underwriting intervention mechanism are triggered in real time, enabling the system to accurately capture the specific risks of battery degradation, system failure, and human-machine conflict.

[0079] (3) Beneficial effects

[0080] Compared with the existing technology, the present invention provides a new energy vehicle insurance dynamic risk pricing analysis system, which has the following beneficial effects:

[0081] 1. In the present invention, by setting up a multi-source acquisition terminal, when conducting new energy vehicle risk analysis, the battery data of the main engine manufacturer, the driving behavior of the Internet of Vehicles and the insurance claim record are obtained in real time, and a multi-source data fusion channel is established to ensure real-time interactive verification of data from different sources; at the same time, the vehicle status is dynamically tracked, and the core risk factors of battery health degradation and driving behavior mutation can be detected in real time, solving the risk mismatch problem caused by the lag of traditional static parameters and improving the timeliness and accuracy of risk identification.

[0082] 2. In the present invention, by setting up a three-dimensional analysis terminal, when conducting insurance actuarial modeling, the coupled analysis of the vehicle condition dimension, driving behavior dimension, and actuarial dimension is performed in parallel, and a five-dimensional factor weight system is constructed based on the entropy weight method; in response to the current situation of multi-source data fragmentation, real-time correlation between battery maintenance records and accident probability is realized, and risk differences of different intelligent driving systems are analyzed in a graded manner, so that the system can eliminate the data island effect. When cross-system data conflicts occur, they are automatically corrected through the weight optimization mechanism to ensure the decision-making accuracy of the actuarial model in a multi-source heterogeneous data environment.

[0083] 3. In the present invention, by setting up a dynamic optimization terminal and a risk warning terminal, when conducting new energy-specific risk management, the battery degradation slope is converted into a dynamic actuarial factor, the contribution of the lidar false detection rate to the accident rate is quantified, and a risk decoupling model for human-machine co-driving is established; when the probability of thermal runaway exceeds the standard, OTA failure, or aggressive driving behavior is identified, a high-risk premium strategy and underwriting intervention mechanism are triggered in real time, allowing the system to accurately capture the specific risks of battery degradation, system failure, and human-machine conflict, solving the risk dimension omissions in traditional models and improving the coverage dimension and risk prevention and control effects of new energy vehicle insurance pricing. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 It is a schematic diagram of the overall system architecture of the present invention. DETAILED DESCRIPTION

[0085] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0086] See also Figure 1 The new energy vehicle insurance dynamic risk pricing analysis system includes a multi-source acquisition terminal, a three-dimensional analysis terminal, a dynamic optimization terminal, a premium calculation terminal, an API service terminal, and a risk warning terminal. The multi-source acquisition terminal, the three-dimensional analysis terminal, the dynamic optimization terminal, the premium calculation terminal, the API service terminal, and the risk warning terminal are jointly provided with a system monitoring module;

[0087] The multi-source acquisition terminal acquires driving data in real time through the vehicle's CAN bus interface, retrieves vehicle configuration information from the OEM, establishes an encrypted connection with the insurance company's database, and integrates user feedback data to form a multi-source data set;

[0088] 3D analysis end execution:

[0089] Vehicle condition analysis: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on historical accident counts and average compensation, and identifies intelligent configuration technology risks.

[0090] Driving behavior dimension analysis: Analyze the proportion of sudden acceleration, sudden braking, and sharp turns, calculate the highway usage rate and the proportion of nighttime driving, and generate behavioral style labels based on intelligent driving logs;

[0091] Actuarial dimension modeling: Construct a five-dimensional factor weight system and calculate the initial weights of industry price factor, historical accident factor, battery factor, intelligent factor, and driving behavior factor using the entropy weight method;

[0092] The dynamic optimization end maps user satisfaction feedback to the corresponding factor dimensions, iteratively updates the factor weight parameters based on the reinforcement learning mechanism, and implements weight lower limit threshold protection for battery factors and risk factors.

[0093] The system also includes:

[0094] The premium calculation end maps the standardized factor values ​​of each dimension to a unified scoring range, calculates the recommended premium value based on dynamic weights, and generates a visual actuarial report that includes the contribution ratio of each factor;

[0095] The API server pushes 3D risk assessment reports to the insurance pricing system through the OAuth2.0 authentication protocol, and simultaneously supports data call requests from the fleet management platform.

[0096] When the product of the historical risk factor scores is greater than 0.25 and the battery health is less than 70%, the risk warning terminal triggers a high-risk flag embedded in the actuarial report, sends a mandatory underwriting instruction to the insurance company, and sets a lower limit for premium output;

[0097] The system monitoring module detects the operating status of each terminal in real time, starts the backup data channel when data anomalies or service interruptions occur, and locates the faulty node through the log analysis engine.

[0098] The system includes the following steps when applied:

[0099] S1. Receive vehicle frame number (VIN) input and invoke a multi-source data acquisition module based on the VIN to obtain real-time and historical vehicle data. This data includes OEM vehicle production data, connected vehicle behavior data, insurance company historical claim data, and user feedback data.

[0100] S2. Parallel execution of 3D risk assessment report generation:

[0101] Vehicle Condition Report: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on the number of historical accidents, types, and average compensation, identifies the intelligence level, and predicts technology iteration risks;

[0102] Driving habit analysis: Analyze driving data to calculate the proportion of sudden acceleration, sudden braking, and sharp turns, the rate of highway use, and the proportion of nighttime driving, and generate behavioral style labels based on the intelligent driving system log;

[0103] Actuarial modeling: Construct a five-dimensional factor weighting system: industry price, historical claims, battery, intelligence, and driving behavior factors, and calculate the initial weights using the entropy weight method;

[0104] S3. Perform dynamic weight optimization:

[0105] Map user satisfaction feedback to corresponding factors;

[0106] Iteratively update factor weights based on human feedback reinforcement learning mechanism, and set time decay coefficient to optimize the impact of recent feedback;

[0107] Set lower weight thresholds for battery factors and risk factors;

[0108] S4. Generate premium actuarial recommendations:

[0109] Standardize and map each factor value to a unified scoring interval;

[0110] Calculate recommended premiums based on dynamic weights and standardized scores, and output a visual report containing the contribution values ​​of each factor;

[0111] S5. Push 3D reports to insurance pricing systems and fleet management platforms through API interfaces.

[0112] The vehicle condition report generation branch in step S2 specifically includes:

[0113] Calculate real-time health based on battery cycle count and voltage fluctuation data:

[0114]

[0115] Where k is the attenuation coefficient, ΔV is the voltage fluctuation amplitude, V max is the maximum nominal voltage of the battery, and T is the cumulative usage time of the battery;

[0116] Build an intelligent configuration risk matrix to identify sensor configurations with a density below the threshold D min 、OTA update interval exceeds T max Model of car;

[0117] LiDAR models and chip models that are at risk of discontinuation are marked based on the supply chain database.

[0118] The behavior style labeling methods in the driving habit analysis branch include:

[0119] Conservative label defined: The intelligent driving system control rate is greater than 70% and the emergency braking ratio is less than 0.1;

[0120] Definition of the aggressive label: the human driving ratio is greater than 60% and the high-speed usage rate is greater than 0.8;

[0121] Define a mixed label: intelligent driving abnormal exit rate ∈ [0.2, 0.5] and night driving ratio ∈ [0.3, 0.6].

[0122] The entropy weight method calculation process includes:

[0123] For n samples, the factor data matrix X m×n Normalization processing:

[0124]

[0125] where x ij is the original value of the jth factor of the i-th sample, minx j 、maxx j is the minimum and maximum value;

[0126] Calculate information entropy:

[0127]

[0128] where p ij =x′ ij 、

[0129] Calculate the initial weights:

[0130]

[0131] in is the initial weight of the j-th factor, is the sum of all factor utility values, 1-E j is the information utility value.

[0132] The feedback mapping rules of the RLHF engine include:

[0133] When the user marks the premium recommendation as "satisfied", the weight of the factor with a contribution value greater than 10% is increased by ΔW = μ·(1-W t ), where μ is the gain coefficient; when marked as “unsatisfactory”, the factor contribution analysis submodule is triggered to identify the score S i Factor ≤ 0.3 and reduce its weight ΔW = -0.1· W t .

[0134] The secondary factors of the five-dimensional factor system include:

[0135] Secondary factor of industry price factor: regional traffic congestion index C I , average annual mileage M avg ;

[0136] Secondary factor of historical accident factor: major accident mark F acc , accident time distribution weight W day / night ;

[0137] Secondary factor of battery factor: battery brand risk coefficient R brand , sensor failure rate λ s ;

[0138] Secondary factor of intelligent factor: ADAS false alarm rate E false , supplier change impact value δ supplier ;

[0139] Secondary factor of driving behavior factor: DMS fatigue driving index I fatigue , extreme weather accident probability P stormo .

[0140] Regional traffic congestion index C I The calculation method is:

[0141]

[0142] Where V real is the real-time vehicle speed, I rush is the peak period coefficient, T is the total number of periods in the statistical period, V max The speed limit of the road.

[0143] The standardized scoring system definition in step S4 is:

[0144] Battery Health Rating:

[0145] S batt =SOH / 100

[0146] Driving behavior score:

[0147]

[0148] Intelligent scoring:

[0149]

[0150] Among them S batt SOH is the battery health score, and SOH is the battery health N accel For rapid acceleration, N brake For emergency braking, N turn is the number of sharp turns, N total is the total number of driving behaviors, L level is the intelligent driving level mapping value, Δt update T is the number of days of OTA delay, std This is the standard upgrade cycle.

[0151] High-risk user warning steps:

[0152] When the condition W is met acc ·S acc ≥0.25, S batt When the value is ≤0.7, an early warning signal is generated and at least one of the following actions is triggered:

[0153] Mark high-risk warning signs in actuarial reports;

[0154] Push mandatory underwriting review requests to the insurance company system;

[0155] Limit the lower limit of the premium output value of the API interface to 120% of the base premium;

[0156] Where W acc is the weight of historical risk factors, S acc Score the risk.

[0157] Implementation 1: Dynamic battery health premium correction:

[0158] In an actual deployment at an insurance company in Zhejiang Province, the system uses a security protocol to obtain real-time battery cycle records, voltage fluctuation curves, and temperature history data from target vehicles. Based on the voltage fluctuation characteristics, it calculates the battery's current health as 88.7% and generates a trend chart showing a monthly average decay rate of 0.25%. When the SOH is detected below the critical value of 90%, the system automatically increases the battery risk factor weight to 0.22. Combined with the 1.32% accident rate in Hangzhou, it dynamically calculates a premium based on the base premium of 2,180 yuan, ultimately outputting a corrected premium of 2,356 yuan. The entire process takes only two hours, an eight-fold increase in efficiency compared to traditional assessments.

[0159] Implementation 2: Decoupling of Dual-Mode Driving Behavior Risks:

[0160] For test vehicles equipped with intelligent driving systems, the system analyzed real-time driving data to separate manual and intelligent driving control records. During the manual driving phase, it detected a 28% excess rate of sudden braking exceeding the safety threshold, and during the intelligent driving phase, a 15% excess rate of abnormal exits. Through dual-channel analysis, it calculated a driving behavior score of 0.62 and an intelligent driving score of 0.4. When a user expressed dissatisfaction with the initial premium feedback, the self-optimization engine dynamically lowered the weight of the driving behavior factor to 0.162 and increased the weight of the intelligent driving factor to 0.236, achieving a 92% accuracy rate in identifying high-risk behaviors.

[0161] Implementation Three: Real-time warning for high-risk users:

[0162] In the taxi fleet monitoring scenario, the system identified that certain vehicles had risk characteristics such as 46% night driving and excessive frequency of high-speed rapid acceleration. Through a composite analysis of the historical accident factor contribution value of 0.105 and the battery health level of 0.68, a three-level early warning mechanism was triggered: a red alert logo was embedded in the actuarial report, a mandatory underwriting instruction was pushed to the insurance company, and a minimum premium guarantee line was set. Within three months of actual operation, the fleet's accident rate dropped by 41%, achieving an annualized premium rationalization saving of 230,000 yuan.

[0163] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0164] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A new energy vehicle insurance dynamic risk pricing analysis system, characterized by: The system includes a multi-source acquisition terminal, a three-dimensional analysis terminal, a dynamic optimization terminal, a premium calculation terminal, an API service terminal, and a risk warning terminal. The multi-source acquisition terminal, the three-dimensional analysis terminal, the dynamic optimization terminal, the premium calculation terminal, the API service terminal, and the risk warning terminal are jointly provided with a system monitoring module; The multi-source acquisition terminal acquires driving data in real time through the vehicle CAN bus interface, retrieves vehicle configuration information from the OEM, establishes an encrypted connection with the insurance company's database, and integrates user feedback data to form a multi-source data set; The three-dimensional analysis terminal performs: Vehicle condition analysis: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on historical accident counts and average compensation, and identifies intelligent configuration technology risks. Driving behavior dimension analysis: Analyze the proportion of sudden acceleration, sudden braking, and sharp turns, calculate the highway usage rate and the proportion of nighttime driving, and generate behavioral style labels based on intelligent driving logs; Actuarial dimension modeling: Construct a five-dimensional factor weight system and calculate the initial weights of industry price factor, historical accident factor, battery factor, intelligent factor, and driving behavior factor using the entropy weight method; The dynamic optimization end maps user satisfaction feedback to corresponding factor dimensions, iteratively updates factor weight parameters based on a reinforcement learning mechanism, and implements weight lower limit threshold protection for battery factors and risk factors.

2. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 1, characterized in that: The system further comprises: The premium calculation terminal maps the standardized factor values ​​of each dimension to a unified scoring range, calculates the recommended premium value based on the dynamic weight, and generates a visual actuarial report including the contribution ratio of each factor; The API server pushes the three-dimensional risk assessment report to the insurance pricing system through the OAuth2.0 authentication protocol, and simultaneously supports data call requests from the fleet management platform; When the product of the historical risk factor scores is greater than 0.25 and the battery health is less than 70%, the risk warning terminal triggers a high-risk flag to be embedded in the actuarial report, sends a mandatory underwriting instruction to the insurance company, and sets a lower limit for premium output; The system monitoring module detects the operating status of each terminal in real time, starts the backup data channel when data is abnormal or service is interrupted, and locates the fault node through the log analysis engine.

3. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 1, characterized in that: The system comprises the following steps: S1. Receive vehicle frame number (VIN) input and invoke a multi-source data acquisition module based on the VIN to obtain real-time and historical vehicle data. The data includes OEM vehicle production data, connected vehicle behavior data, insurance company historical claim data, and user feedback data. S2. Parallel execution of 3D risk assessment report generation: Vehicle Condition Report: Calculates the battery health status (SOH) value and generates a decay trend chart, compiles statistics on the number of historical accidents, types, and average compensation, identifies the intelligence level, and predicts technology iteration risks; Driving habit analysis: Analyze driving data to calculate the proportion of sudden acceleration, sudden braking, and sharp turns, the rate of highway use, and the proportion of nighttime driving, and generate behavioral style labels based on the intelligent driving system log; Actuarial modeling: Construct a five-dimensional factor weighting system: industry price, historical claims, battery, intelligence, and driving behavior factors, and calculate the initial weights using the entropy weight method; S3. Perform dynamic weight optimization: Map user satisfaction feedback to corresponding factors; Iteratively update factor weights based on human feedback reinforcement learning mechanism, and set time decay coefficient to optimize the impact of recent feedback; Set lower weight thresholds for battery factors and risk factors; S4. Generate premium actuarial recommendations: Standardize and map each factor value to a unified scoring interval; Calculate recommended premiums based on dynamic weights and standardized scores, and output a visual report containing the contribution values ​​of each factor; S5. Push 3D reports to insurance pricing systems and fleet management platforms through API interfaces.

4. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 3, characterized in that: The vehicle condition report generation branch in step S2 specifically includes: Calculate real-time health based on battery cycle count and voltage fluctuation data: Where SOH is the battery health, k is the attenuation coefficient, ΔV is the voltage fluctuation amplitude, V max is the maximum nominal voltage of the battery, and T is the cumulative usage time of the battery; Build an intelligent configuration risk matrix to identify sensor configurations with a density below the threshold D min 、OTA update interval exceeds T max Model of car; LiDAR models and chip models that are at risk of discontinuation are marked based on the supply chain database.

5. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 3, characterized in that: The behavior style labeling method in the driving habit analysis branch includes: Conservative label defined: The intelligent driving system control rate is greater than 70% and the emergency braking ratio is less than 0.1; Definition of the aggressive label: the human driving ratio is greater than 60% and the high-speed usage rate is greater than 0.8; Define a mixed label: intelligent driving abnormal exit rate ∈ [0.2, 0.5] and night driving ratio ∈ [0.3, 0.6].

6. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 3, characterized in that: The entropy weight method calculation process includes: For n samples, the factor data matrix X m×n Normalization processing: where x ij is the original value of the jth factor of the i-th sample, is the minimum and maximum value; Calculate information entropy: in Calculate the initial weights: in is the initial weight of the j-th factor, is the sum of all factor utility values, 1-E j is the information utility value.

7. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 3, characterized in that: The feedback mapping rules of the RLHF engine include: When the user marks the premium recommendation as "satisfied", the weight of the factor with a contribution value greater than 10% is increased by ΔW = μ·(1-W t ), where μ is the gain coefficient; when marked as "unsatisfactory", the factor contribution analysis submodule is triggered to identify the score S i Factor ≤0.3 and reduce its weight ΔW=-0.1·W t .

8. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 3, characterized in that: The secondary factors of the five-dimensional factor system include: Secondary factor of industry price factor: regional traffic congestion index C I , average annual mileage M avg ; Secondary factor of historical accident factor: major accident mark F acc , accident time distribution weight W day / night ; Secondary factor of battery factor: battery brand risk coefficient R brand , sensor failure rate λ s ; Secondary factor of intelligent factor: ADAS false alarm rate E false , supplier change impact value δ supplier ; Secondary factor of driving behavior factor: DMS fatigue driving index I fatigue , extreme weather accident probability P stormo .

9. A new energy vehicle insurance dynamic risk pricing analysis system according to claim 8, characterized in that: Traffic congestion index C of the area I The calculation method is: Where V real is the real-time vehicle speed, I rush is the peak period coefficient, T is the total number of periods in the statistical period, V max The speed limit of the road.

10. Based on a new energy vehicle insurance dynamic risk pricing analysis system, it is characterized by: The standardized scoring system in step S4 is defined as: Battery Health Rating: S batt =SOH / 100 Driving behavior score: Intelligent scoring: Among them S batt SOH is the battery health score, N accel For rapid acceleration, N brake For emergency braking, N turn is the number of sharp turns, N total is the total number of driving behaviors, L level is the intelligent driving level mapping value, Δt update T is the number of days of OTA delay, std This is the standard upgrade cycle; High-risk user warning steps: When the condition W is met acc ·S acc ≥0.25, S batt When the value is ≤0.7, an early warning signal is generated and at least one of the following actions is triggered: Mark high-risk warning signs in actuarial reports; Push mandatory underwriting review requests to the insurance company system; Limit the lower limit of the premium output value of the API interface to 120% of the base premium; Where W acc is the weight of historical risk factors, S acc Score the risk.

Images