Multi-dimensional data fusion system for operating truck risk rating

Abstract

The invention provides a multi-dimensional data fusion system for operating truck risk rating, and relates to the technical field of data processing, and the system comprises a data collection module which is used for obtaining the dynamic operation data and historical static information of a target operating truck; the data preprocessing module is used for performing field alignment, abnormal value elimination and format standardization; the feature analysis module is used for extracting dynamic behavior features of vehicle operation and performing statistical calculation in a preset time window; the event identification and risk assignment module is used for identifying a single event or a combined event, assigning an event risk weight to the single event or the combined event, and determining a corresponding road condition risk level in combination with the vehicle position information; the multi-dimensional fusion module is used for fusing features and generating time serialized multi-dimensional risk feature vectors; the risk rating module is used for scoring the multi-dimensional risk feature vector and outputting a corresponding risk rating result; according to the invention, the accuracy of the multi-dimensional data fusion system is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a multi-dimensional data fusion system for risk rating of operating trucks. Background Art

[0002] Existing risk rating technologies for commercial trucks often rely on single-source data, such as a vehicle's historical violation records, annual inspection results, or a single GPS track. These systems typically assign weights to these single-dimensional data and match them with pre-established risk models to generate a risk score. For example, some fleet management platforms generate risk ratings based solely on a driver's past traffic violations and annual vehicle inspection results, combined with simple scoring rules. This approach is cost-effective and simple, but its limited data dimensions make it difficult to fully reflect the risk profile of a vehicle under different operating conditions.

[0003] In interprovincial, long-distance transport scenarios, existing rating methods that rely on single-dimensional data can risk missing detections. For example, a truck may have no major local violations and pass annual inspections, but may experience repeated rapid acceleration and braking while driving continuously on the highway. These behaviors can only be captured through dynamic sensor data during vehicle operation. Because existing systems don't incorporate this real-time, dynamic data, the truck's risk rating may still be low, misleading decisions in insurance rate assessments or dispatch allocations, leading to potential safety hazards being overlooked. Summary of the Invention

[0004] The purpose of the present invention is to provide a multi-dimensional data fusion system for operating truck risk rating, aiming to solve the problems mentioned in the background technology.

[0005] In order to solve the above technical problems, the technical solutions of the present invention are as follows: A multi-dimensional data fusion system for risk rating of operating trucks, the system comprising: The data acquisition module is used to obtain the dynamic operation data and historical static information of the target operating trucks to form a preliminary multi-dimensional data set; The data preprocessing module is used to align fields, remove outliers, and normalize the format of the preliminary multidimensional data set to generate structured multidimensional data; The feature analysis module is used to extract the dynamic behavior characteristics of vehicle operation from structured multidimensional data and perform statistical calculations within a preset time window to form a feature statistical value data set; The event identification and risk assignment module is used to identify single events or combined events based on the characteristic statistical value data set and preset event trigger conditions, and assign event risk weights to them. At the same time, it combines the vehicle location information to determine the corresponding road condition risk level and summarizes it to form an event risk feature set; The multidimensional fusion module is used to fuse the event risk feature set with the historical static information in the structured multidimensional data to generate a time-series multidimensional risk feature vector; The risk rating module is used to score the multi-dimensional risk feature vector and output the corresponding risk rating results.

[0006] Preferably, the feature analysis module includes: A dynamic feature segmentation unit is used to divide the dynamic operation data into multiple time periods according to a preset time window to obtain time segmentation data; The characteristic value calculation unit is used to process the time-segmented data, calculate the instantaneous value and cumulative value of the dynamic behavior characteristics, and form a characteristic value set; The statistics generation unit is used to calculate the change amplitude, peak value and cumulative impact of each dynamic behavior feature within a time period based on the feature value set, and generate a feature statistical value data set.

[0007] Preferably, the event identification and risk assessment module includes: An event identification unit is used to identify a single event or a combination of events of a target operating truck within a preset time period based on a feature statistical value data set and preset trigger conditions to obtain an event set; The weight allocation unit is used to assign a preset risk weight value to a single event according to the event type of the event set, and perform a weighted operation on the statistical value of each dynamic behavior feature of the combined event and the corresponding risk weight to obtain the event risk weight value; The road condition assessment unit is used to combine the location information of the event set, search the preset road condition risk level table, determine the corresponding road condition risk level, and summarize it with the road condition risk level and event risk weight value to form an event risk feature set.

[0008] Preferably, the multi-dimensional fusion module includes: A feature matching unit is used to associate and match the event risk feature set with the historical static information of the target operating truck based on the unique identifier of the target operating truck and the timestamp of the event, to obtain fused input data corresponding to the event; The weighted fusion unit is used to multiply the event risk feature set and the historical static information parameters by the preset weighting coefficients according to the fusion input data, and then sum them up to generate a multidimensional risk feature vector containing static factors and dynamic factors; The sequence generating unit is used to arrange the multi-dimensional risk feature vectors in chronological order to form a multi-dimensional risk feature vector sequence.

[0009] Preferably, the statistics generating unit includes: An amplitude calculation unit, configured to calculate the difference between the maximum value and the minimum value of each dynamic behavior feature within a time period based on the feature value set, and generate an amplitude calculation result; The peak extraction unit is used to analyze the time series of dynamic behavior characteristics based on the amplitude calculation results, identify local maximum values ​​and record their size and occurrence time to generate peak extraction results; The cumulative statistical unit is used to accumulate all dynamic behavior feature values ​​within the time period item by item based on the peak extraction result, and weight the values ​​of different time segments with preset weights to form a cumulative impact statistical result; The result generating unit is used to combine the amplitude calculation result, the peak value extraction result and the cumulative impact quantity statistical result to generate a characteristic statistical value data set.

[0010] Preferably, the event determination unit includes: a single event detection unit, configured to compare the statistical value of each dynamic behavior feature with the corresponding preset trigger condition based on the feature statistical value data set, and generate a single event record when the statistical value of a dynamic behavior feature exceeds a preset first number threshold; a combined event identification unit, configured to analyze multiple single events within the same time window based on the single event record, and combine the single events into a combined event when a preset combination condition is met, thereby generating a combined event record, wherein the preset combination condition is that different types of single events are triggered simultaneously within the same time window, or that the number of times the same type of single event is triggered exceeds a preset second number threshold within a preset time interval; The event set generating unit is configured to aggregate the single event record and the combined event record to generate an event set.

[0011] Preferably, the weight distribution unit includes: An initial weight generating unit, configured to determine an initial risk weight value based on the event type of the event set and by comparing it with a preset risk weight table; A dynamic adjustment unit is used to modify the initial risk weight value based on historical accident data and operating environment data to generate a dynamically adjusted event risk weight value. The modification process is as follows: the risk level corresponding to different types of events is multiplied by their occurrence frequency in historical accident samples, and all product results are normalized; The nonlinear interaction unit is used to process the combined events according to the dynamically adjusted event risk weight value. When two or more dynamic behavior features are triggered simultaneously in the same time window, the statistical value of each feature is multiplied by the corresponding dynamically adjusted weight value to obtain a single risk contribution. Then, multiple risk contributions are enhanced and superimposed according to the preset amplification factor to generate a combined event weight value.

[0012] Preferably, the road condition assessment unit includes: A basic level retrieval unit is used to compare the location information of the target operating truck with a preset road condition risk level table to determine the initial road condition risk level; An external data acquisition unit is used to obtain external environmental data of the road section where the vehicle is located when an incident occurs, wherein the external environmental data includes the degree of road congestion, real-time weather conditions and historical accident frequency, and generate external road condition data; The level correction unit is used to correct the initial road condition risk level according to the external road condition data. When the external environment data indicates that the road risk increases, the risk level is increased. When the external environment data indicates that the road risk decreases, the risk level is reduced to obtain a dynamically updated road condition risk level.

[0013] Preferably, the weighted fusion unit includes: A coefficient optimization unit is used to iteratively optimize preset weighting coefficients by training multi-dimensional data samples of historical operating trucks to generate optimized weighting coefficients. The iterative optimization process includes: inputting historical samples into the machine learning model, comparing the risk feature vectors output by the model with the corresponding accident records, and gradually adjusting the weighting coefficients until convergence; The time series weighting unit is used to fuse the event risk feature set and historical static information parameters in the fusion input data according to the optimized weighting coefficient. In the processing, a time decay factor is introduced to reduce the weight of data at earlier moments and increase the weight of recent data to obtain the time series weighted result. The nonlinear fusion unit is used to perform nonlinear mapping and weighted summation on the interaction between the event risk feature set and historical static information based on the time series weighted results, and generate a multidimensional risk feature vector containing static factors and dynamic factors.

[0014] Preferably, the nonlinear fusion unit includes: The feature interaction modeling unit is used to model the feature interaction relationship between the event risk feature set and historical static information based on the time series weighted results, convert the interaction strength of the two types of features into interaction feature values, and form the interaction modeling results; A nonlinear mapping unit is used to perform nonlinear function transformation on the dynamic behavior characteristics, event risk weights, and road condition risk levels based on the interactive modeling results, and to combine the mapping values ​​in a layer-by-layer iterative manner to generate a nonlinear mapping result, wherein the nonlinear function is a piecewise function that can amplify the contribution of high-risk features and suppress the noise of low-risk features; The fusion generation unit is used to perform weighted summation of dynamic behavior characteristics and static information parameters according to the nonlinear mapping results to obtain a multidimensional risk feature vector containing feature interactions.

[0015] The above solution of the present invention includes at least the following beneficial effects: By integrating a data acquisition module, a data preprocessing module, a feature analysis module, an event identification and risk assessment module, a multi-dimensional fusion module, and a risk rating module, a comprehensive multi-dimensional assessment of operational truck risk ratings can be achieved. Compared to existing solutions that rely on a single data source, this invention simultaneously captures both dynamic operational data and historical static information about vehicles and processes them within a unified structured data framework, ensuring the comprehensiveness and reliability of the data used in the risk analysis process.

[0016] At the data processing level, the feature analysis module extracts dynamic behavior characteristics such as acceleration, braking intensity, and steering amplitude within a preset time window, and calculates their amplitude, peak value, and cumulative impact, thereby truly reflecting the vehicle's actual driving behavior. This processing approach avoids the problem of inaccurate risk assessment caused by relying solely on static information such as violation records and annual inspection results.

[0017] At the risk event level, the event recognition and risk assignment module transforms abstract numerical features into identifiable driving events, combines location information with road risk levels, and ultimately aggregates them into a set of event risk signatures. For example, in interprovincial highway transport scenarios, if a vehicle frequently exhibits combined behaviors such as sudden acceleration and sharp turns, this invention can identify it as a high-risk event, thus preventing traditional methods from misclassifying it as a low-risk vehicle.

[0018] At the fusion analysis level, this invention uses a multidimensional fusion module to perform a weighted fusion of event risk signatures with the vehicle's historical static information, generating a time-series multidimensional risk signature vector. This approach not only preserves the vehicle's basic static risk information but also overlays dynamic risk signatures generated during operation, providing a comprehensive description of the vehicle's risk status.

[0019] The risk rating results ultimately output by the risk rating module can be used to intuitively reflect the vehicle's risk level in the form of a score or grade, facilitating direct application in scenarios such as insurance rate assessment, fleet scheduling optimization, and transportation safety supervision. This invention effectively addresses the existing issues of missed and misjudgment of risks caused by limited data dimensions, improving the accuracy and practicality of risk rating results and reducing potential safety hazards. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 11 is an architectural diagram of a multi-dimensional data fusion system for operating truck risk rating provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0021] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0022] like Figure 1 As shown, an embodiment of the present invention proposes a multi-dimensional data fusion system for operating truck risk rating, the system comprising: The data acquisition module is used to obtain the dynamic operation data and historical static information of the target operating trucks to form a preliminary multi-dimensional data set; The data preprocessing module is used to align fields, remove outliers, and normalize the format of the preliminary multidimensional data set to generate structured multidimensional data; The feature analysis module is used to extract the dynamic behavior characteristics of vehicle operation from structured multidimensional data and perform statistical calculations within a preset time window to form a feature statistical value data set; The event identification and risk assignment module is used to identify single events or combined events based on the characteristic statistical value data set and preset event trigger conditions, and assign event risk weights to them. At the same time, it combines the vehicle location information to determine the corresponding road condition risk level and summarizes it to form an event risk feature set; The multidimensional fusion module is used to fuse the event risk feature set with the historical static information in the structured multidimensional data to generate a time-series multidimensional risk feature vector; The risk rating module is used to score the multi-dimensional risk feature vector and output the corresponding risk rating results.

[0023] In this embodiment of the present invention, a comprehensive multi-dimensional assessment of operational risks for commercial trucks is achieved by configuring a data acquisition module, a data preprocessing module, a feature analysis module, an event identification and risk assignment module, a multi-dimensional fusion module, and a risk rating module. The system simultaneously acquires dynamic operational data and historical static information from vehicles, generating structured multi-dimensional data after data preprocessing, thereby ensuring the reliability and accuracy of subsequent analysis. Feature analysis extracts dynamic behavioral characteristics such as vehicle acceleration, braking intensity, and steering amplitude within a preset time window, and calculates their amplitude, peak value, and cumulative impact, thus avoiding the problem of inaccurate risk assessment caused by reliance on a single data source.

[0024] In practice, the event identification and risk assessment module identifies single or combined events, assigns them risk weights, and combines them with vehicle location information to determine the road risk level. Ultimately, this module aggregates these information into a set of event risk signatures. For example, during long-distance highway transport, if a truck frequently engages in a combination of sudden acceleration and sharp turns despite having no record of major violations, the system can accurately reflect its potential risk using the event risk signature set, rather than relying solely on historical static data.

[0025] The multidimensional fusion module performs a weighted fusion of event risk signatures and historical static vehicle information to generate a time-series multidimensional risk feature vector. This allows risk representation to encompass not only dynamic behavior but also static information such as annual vehicle inspection results and past violations, thereby achieving deep integration of multi-source data. Based on this, the risk rating module outputs a quantitative risk rating result in the form of a score or grade, facilitating direct application in insurance pricing, fleet scheduling, or safety supervision.

[0026] Therefore, the system can comprehensively and dynamically evaluate the risk level of operating trucks in combination with the specific operating environment, effectively solving the problem of missed or misjudgment of risks caused by existing reliance on single-dimensional data, and improving the accuracy and guidance of risk rating results in practical applications.

[0027] In one embodiment of the present invention, the dynamic operating data may include, but is not limited to, vehicle speed, longitudinal acceleration, lateral acceleration, braking intensity, fuel consumption rate, engine speed, steering wheel angle, driving time, and continuous driving time; the historical static information may include, but is not limited to, vehicle annual inspection records, repair and maintenance records, historical insurance claims records, traffic violation records, vehicle production date and age, etc. The above data can be obtained through vehicle-mounted terminal devices, vehicle sensors, telematics platforms, and data interfaces of traffic management departments.

[0028] In a specific application scenario, risk rating results can be expressed as a numerical value on a scale of 0–100, with higher scores indicating greater risk. Alternatively, they can be expressed as a tiered scale, such as 1–5, with 1 indicating low risk and 5 indicating high risk. This rating can directly inform insurance pricing, with higher insurance premiums assigned to high-risk vehicles. It can also be used in fleet scheduling and operations management, such as prioritizing high-risk vehicles for short-distance transport or low-risk routes. This design enables the system to integrate dynamic and static information for analysis, ensuring the scientific nature and practical value of risk rating results.

[0029] In one embodiment of the present invention, the dynamic operation data collected by the data acquisition module can be acquired at a sampling frequency of 10Hz to 100Hz to ensure real-time tracking of the vehicle's operating status. Among them, high-frequency changing features such as acceleration and braking intensity are preferably sampled at 50Hz, and low-frequency features such as steering wheel angle can be sampled at 10Hz to reduce data redundancy. In order to ensure data quality, the data preprocessing module can eliminate abnormal values ​​during the acquisition process. When the vehicle acceleration exceeds ±20 meters per square second or the fuel consumption rate exceeds a reasonable range, it is automatically marked as abnormal and eliminated. In addition, the data from different sensors are aligned by time stamps and uniformly interpolated to the same sampling frequency, thereby forming structured multidimensional data with a consistent format to ensure the reliability of subsequent feature analysis links.

[0030] In a preferred embodiment of the present invention, the feature analysis module includes: A dynamic feature segmentation unit is used to divide the dynamic operation data into multiple time periods according to a preset time window to obtain time segmentation data; The characteristic value calculation unit is used to process the time-segmented data, calculate the instantaneous value and cumulative value of the dynamic behavior characteristics, and form a characteristic value set; The statistics generation unit is used to calculate the change amplitude, peak value and cumulative impact of each dynamic behavior feature within a time period based on the feature value set, and generate a feature statistical value data set.

[0031] In an embodiment of the present invention, by setting a dynamic feature segmentation unit, a feature value calculation unit and a statistical generation unit in the feature analysis module, segmented processing and feature quantification of vehicle operation data are achieved. The dynamic feature segmentation unit can divide the dynamic operation data of the vehicle into multiple time periods according to a preset time window, so that the system can observe the vehicle operation characteristics in a fine-grained time dimension. On this basis, the feature value calculation unit extracts the instantaneous value and cumulative value in each time period to ensure that it can capture both sudden behaviors and long-term cumulative effects. The statistical generation unit further calculates the amplitude, peak value and cumulative impact of these feature values ​​to obtain a structured feature statistical value data set.

[0032] Through this processing, the system can establish a connection between short-term driving behaviors and longer-term driving habits. For example, if a truck frequently exceeds the acceleration limit and its cumulative braking intensity is far above average during an hour's driving, even if the vehicle has a good history of violations, the module's analysis can identify its potential high-risk driving behavior, thereby improving the accuracy of subsequent event recognition.

[0033] In one embodiment of the present invention, the preset time window can be set to a fixed duration (e.g., 5 minutes, 10 minutes, or 30 minutes) according to actual needs, or it can be set to a fixed mileage segment (e.g., 5 kilometers or 10 kilometers) according to the vehicle's cumulative mileage. The time window length can also be dynamically adjusted through an algorithm to adapt to different operating environments.

[0034] Within this time window, the characteristic value calculation unit calculates the instantaneous and cumulative values ​​of the dynamic behavior characteristics. The instantaneous value can be defined as the instantaneous physical quantity at the sampling moment, such as the longitudinal acceleration at a specific moment. The cumulative value can be obtained by adding up or averaging all the instantaneous values ​​within the time window to reflect the overall level of driving behavior. For example, if 60 longitudinal acceleration values ​​are collected within a 10-minute time window, the instantaneous value is the value at each sampling time, and the cumulative value can be the sum of these 60 values ​​or the average value obtained by dividing the sum by the number of sampling times.

[0035] In a specific application scenario, if a truck experiences instantaneous longitudinal acceleration exceeding 3 meters per square second multiple times within a 10-minute time window, and the cumulative braking intensity within that time window is much higher than the fleet average, the characteristic statistical value dataset generated by the statistical generation unit will clearly highlight the risk characteristics of the vehicle, providing high-quality input data for subsequent event identification.

[0036] In one embodiment of the present invention, the feature analysis module employs different calculation methods for different types of dynamic behavior features when performing feature calculations. The cumulative value of the vehicle's longitudinal acceleration can be obtained by cumulatively adding up the instantaneous values ​​within a time window. The cumulative value of braking intensity can be obtained by weighted accumulation of the absolute values ​​of deceleration. The cumulative value of steering amplitude can be obtained by cumulatively adding up the absolute changes in steering wheel angles to reflect the frequency of steering.

[0037] In a preferred embodiment of the present invention, the event identification and risk assessment module includes: An event identification unit is used to identify a single event or a combination of events of a target operating truck within a preset time period based on a feature statistical value data set and preset trigger conditions to obtain an event set; The weight allocation unit is used to assign a preset risk weight value to a single event according to the event type of the event set, and perform a weighted operation on the statistical value of each dynamic behavior feature of the combined event and the corresponding risk weight to obtain the event risk weight value; The road condition assessment unit is used to combine the location information of the event set, search the preset road condition risk level table, determine the corresponding road condition risk level, and summarize it with the road condition risk level and event risk weight value to form an event risk feature set.

[0038] In an embodiment of the present invention, by setting an event discrimination unit, a weight allocation unit and a road condition assessment unit in the event identification and risk assignment module, the risk event representation of dynamic behavior characteristics is achieved. The event discrimination unit can identify a single event or a combination of events based on the characteristic statistical value data set, thereby converting abstract numerical features into specific driving event records. On this basis, the weight allocation unit assigns preset risk weight values ​​to different types of events, and obtains event risk weights through weighted calculation, so that the contribution of different events to the overall risk can be quantified. The road condition assessment unit combines the vehicle location information, searches the road condition risk level table, modifies the event risk weight value and outputs the event risk feature set.

[0039] In this way, the system not only identifies the inherent risk of driving behavior but also adjusts the risk level based on the external environment. For example, the same sudden braking incident might be assigned a lower risk level if it occurs on a city road, but the risk level would be significantly higher if it occurs on a rainy highway. This module dynamically adjusts the risk representation of events in different scenarios, thus avoiding missed or misjudgment caused by a single risk weight.

[0040] In one embodiment of the present invention, the preset event triggering conditions can be set based on traffic regulations, industry safety standards, and historical accident statistics. For example, a sudden acceleration event is determined when the instantaneous longitudinal acceleration of the vehicle is greater than 3 meters per second squared, a sudden braking event is determined when the braking deceleration is less than -4 meters per second squared, and a sharp turn event is determined when the steering wheel angular velocity is greater than 300 degrees per second.

[0041] In this embodiment, the event identification unit analyzes the feature statistics dataset to identify single or combined events within a preset time period for the target truck. For example, if a sudden acceleration and a sharp turn are detected within the same time window, it can be identified as a "dangerous overtaking event." If three sudden braking events occur consecutively within a five-minute period, it can be identified as a "high-risk fatigue driving event."

[0042] The weight assignment unit assigns a preset risk weight to each event based on the event type within the event set, such as 0.3 for a sudden braking event, 0.2 for a sudden acceleration event, and 0.25 for a sudden turn event. For combined events, the risk weight is calculated by multiplying the statistical value of each feature by the corresponding weight and summing the results. For example, for a "sudden acceleration + sudden turn" event, the resulting weight is 0.2 × acceleration statistic + 0.25 × turn statistic.

[0043] The road condition assessment unit combines the location information of the event set with the corresponding level in a pre-set road condition risk level table. For example, the risk level for expressways is 1, for national highways is 2, and for mountain roads is 3. In a specific scenario, if a sudden braking event occurs on a rainy mountain road, the road condition assessment unit will significantly increase the risk level. This revised risk level is then combined with the event risk weight to form an event risk signature set. This approach avoids the omissions and misjudgments that can occur with traditional risk assessments based on a single piece of data.

[0044] In a preferred embodiment of the present invention, the multi-dimensional fusion module includes: A feature matching unit is used to associate and match the event risk feature set with the historical static information of the target operating truck based on the unique identifier of the target operating truck and the timestamp of the event, to obtain fused input data corresponding to the event; The weighted fusion unit is used to multiply the event risk feature set and the historical static information parameters by the preset weighting coefficients according to the fusion input data, and then sum them up to generate a multidimensional risk feature vector containing static factors and dynamic factors; The sequence generating unit is used to arrange the multi-dimensional risk feature vectors in chronological order to form a multi-dimensional risk feature vector sequence.

[0045] In an embodiment of the present invention, a feature matching unit, a weighted fusion unit, and a sequence generation unit are incorporated into the multidimensional fusion module to achieve a deep fusion of event risk feature sets and historical static information. The feature matching unit can associate real-time event risk features with the vehicle's historical records based on the vehicle's unique identifier and event timestamp, freeing risk analysis from being limited to a single dimension. The weighted fusion unit then performs a weighted summation of the event risk feature set with static parameters such as vehicle annual inspection results and violation records to produce a multidimensional risk feature vector encompassing both dynamic and static factors. The sequence generation unit further arranges these vectors in chronological order to form a time-series input.

[0046] The technical benefit of this fusion approach lies in its ability to complement dynamic and static risk information. For example, if a vehicle has a minor failure in its annual inspection, but has recently experienced high-risk events such as sharp turns, the fused multidimensional risk feature vector will significantly increase the vehicle's overall risk level, making the rating more consistent with its actual operating conditions. In contrast, traditional rating methods based on a single static piece of information may underestimate the vehicle's risk level. However, the fusion mechanism of this invention effectively improves the accuracy and sensitivity of risk assessment.

[0047] In one embodiment of the present invention, the feature matching unit can match the event risk feature set with the historical static information of the target operating truck based on the unique identifier of the target operating truck (such as the license plate number or vehicle identification number (VIN)) and the timestamp of the event. For example, if a vehicle experiences a sudden braking event at 10:00 on January 10, 2025, the system can correlate this event with the vehicle's historical violation records, annual inspection results, and insurance claims records to generate fused input data.

[0048] The weighted fusion unit multiplies the event risk signature set and historical static information parameters by preset weighting coefficients and then takes a weighted sum to form a risk signature vector containing multiple factors. For example, in a calculation scenario, if the event risk signature set accounts for 0.5, historical violation records account for 0.3, and vehicle annual inspection information accounts for 0.2, the final risk signature vector can be expressed as [0.5 × event risk signature value + 0.3 × violation signature value + 0.2 × annual inspection signature value].

[0049] The sequence generation unit can arrange the multidimensional risk feature vectors generated above in chronological order to form a multidimensional risk feature vector sequence. This way, the risk status at different moments is preserved and can be provided as input to the risk rating module. For example, during a day's transportation, a vehicle may experience a variety of scenarios, including low risk (normal driving on urban roads), medium risk (sudden braking on highways), and high risk (sharp turns on mountain roads). The sequence generation unit can string together the risk characteristics of these different time periods into a complete risk change trajectory, which can be used to more accurately determine the risk score.

[0050] In a preferred embodiment of the present invention, the statistics generating unit includes: An amplitude calculation unit, configured to calculate the difference between the maximum value and the minimum value of each dynamic behavior feature within a time period based on the feature value set, and generate an amplitude calculation result; The peak extraction unit is used to analyze the time series of dynamic behavior characteristics based on the amplitude calculation results, identify local maximum values ​​and record their size and occurrence time to generate peak extraction results; The cumulative statistical unit is used to accumulate all dynamic behavior feature values ​​within the time period item by item based on the peak extraction result, and weight the values ​​of different time segments with preset weights to form a cumulative impact statistical result; The result generating unit is used to combine the amplitude calculation result, the peak value extraction result and the cumulative impact quantity statistical result to generate a characteristic statistical value data set.

[0051] In this embodiment of the present invention, by arranging an amplitude calculation unit, a peak extraction unit, a cumulative statistics unit, and a result generation unit within a statistics generation unit, multi-angle quantitative processing of dynamic behavior characteristics is achieved. The amplitude calculation unit can reflect the fluctuation range of the characteristic within a time window by the difference between the maximum and minimum values. The peak extraction unit can capture extreme driving maneuvers. The cumulative statistics unit can reflect the cumulative impact of driving behavior throughout the entire window. The result generation unit integrates the results of these different dimensions to form a characteristic statistical value dataset.

[0052] This multi-level feature extraction approach allows the system to comprehensively reflect the risks involved in vehicle operation. For example, if two trucks both brake suddenly during the same time period, but one exhibits a greater range of variation, a higher peak intensity, and a significantly larger cumulative impact, the statistical generation unit will highlight the risk characteristics of this vehicle, providing more accurate data support for subsequent event identification, thereby improving the discriminability of risk ratings.

[0053] In one embodiment of the present invention, the amplitude calculation unit can compare and subtract the maximum and minimum values ​​of each dynamic behavior feature in the feature value set within a time window to obtain the amplitude of change of the feature. For example, if the maximum longitudinal acceleration of the vehicle is 4 meters per second squared and the minimum is -2 meters per second squared within a 10-minute time window, the calculated acceleration amplitude is 6 meters per second squared.

[0054] The peak extraction unit identifies local maxima or minima in a time series and records their magnitude and the time of their occurrence. For example, if a vehicle experiences three peaks in braking intensity within a 10-minute window—–5, –6, and –7 meters per second squared—the system will record these peaks and their corresponding times.

[0055] The cumulative statistics unit accumulates all dynamic behavior feature values ​​within a time window and weights different time segments according to preset weights to highlight the risk impact of recent behavior. For example, if the vehicle acceleration is sampled 100 times in a window, the cumulative statistics unit can add up all 100 values ​​and assign a higher weight to the last 20 data points in the window to reflect the importance of recent driving behavior.

[0056] The result generation unit combines the amplitude calculation results, peak extraction results, and cumulative impact statistics to generate a feature statistical value dataset. This dataset comprehensively reflects the vehicle's instantaneous risk, extreme operation risk, and long-term cumulative risk, providing high-quality input data for subsequent event identification.

[0057] In a preferred embodiment of the present invention, the event determination unit includes: a single event detection unit, configured to compare the statistical value of each dynamic behavior feature with the corresponding preset trigger condition based on the feature statistical value data set, and generate a single event record when the statistical value of a dynamic behavior feature exceeds a preset first number threshold; a combined event identification unit, configured to analyze multiple single events within the same time window based on the single event record, and combine the single events into a combined event when a preset combination condition is met, thereby generating a combined event record, wherein the preset combination condition is that different types of single events are triggered simultaneously within the same time window, or that the number of times the same type of single event is triggered exceeds a preset second number threshold within a preset time interval; The event set generating unit is configured to aggregate the single event record and the combined event record to generate an event set.

[0058] In this embodiment of the present invention, by incorporating a single event detection unit, a combined event identification unit, and an event set generation unit within the event discrimination unit, hierarchical identification and set management of driving events is achieved. The single event detection unit can identify single risk behaviors such as sudden braking and acceleration based on characteristic statistics and preset thresholds. The combined event identification unit further determines whether multiple single events are triggered concurrently within the same time window, or whether similar events occur frequently within a certain time interval, and generates a combined event record when preset combination conditions are met. The event set generation unit can integrate single events and combined events into an event set.

[0059] This design enables the system to more accurately characterize complex driving behaviors. For example, a single sudden braking event might represent an emergency, but if it occurs repeatedly within a short period of time, it could indicate fatigue. Similarly, if sudden acceleration and sharp turns occur simultaneously within the same window, it could indicate dangerous overtaking behavior. By identifying and collectively managing combined events, the system can more accurately reflect the potential risks of driving behavior, avoiding omissions and underestimations caused by single-event judgments.

[0060] In one embodiment of the present invention, the single event detection unit compares characteristic statistics with a preset threshold and generates a single event record when a dynamic behavior characteristic exceeds the threshold. For example, if the vehicle's longitudinal acceleration exceeds 3 meters per second squared within 5 minutes, it is determined to be a "sudden acceleration event"; if the braking deceleration is less than -4 meters per second squared, it is determined to be a "sudden braking event."

[0061] The combined event recognition unit analyzes single event records and generates a combined event record when multiple single events of different types are triggered simultaneously within the same time window, or when similar events occur frequently within a certain time interval. For example, detecting both sudden acceleration and a sharp turn within the same time window could be considered a "dangerous overtaking event," while three consecutive sudden braking events within 10 minutes could be considered a "high-risk fatigue driving event."

[0062] The event set generation unit integrates single and combined events into a unified event set. This event set not only represents a single risk event within a specific time window, but also reflects complex driving risk situations involving multiple events. For example, if a truck driving on a mountain road triggers both a sudden braking and a sharp turn, the event set will retain this combined event, enabling more accurate risk weighting and road condition correction.

[0063] In a preferred embodiment of the present invention, the weight allocation unit includes: An initial weight generating unit, configured to determine an initial risk weight value based on the event type of the event set and by comparing it with a preset risk weight table; A dynamic adjustment unit is used to modify the initial risk weight value based on historical accident data and operating environment data to generate a dynamically adjusted event risk weight value. The modification process is as follows: the risk level corresponding to different types of events is multiplied by their occurrence frequency in historical accident samples, and all product results are normalized; The nonlinear interaction unit is used to process the combined events according to the dynamically adjusted event risk weight value. When two or more dynamic behavior features are triggered simultaneously in the same time window, the statistical value of each feature is multiplied by the corresponding dynamically adjusted weight value to obtain a single risk contribution. Then, multiple risk contributions are enhanced and superimposed according to the preset amplification factor to generate a combined event weight value.

[0064] In an embodiment of the present invention, by providing an initial weight generation unit, a dynamic adjustment unit, and a nonlinear interaction unit in the weight allocation unit, hierarchical weighting and dynamic correction of risk events are achieved. The initial weight generation unit can determine the basic risk weight of the event based on the event type and a preset risk weight table, while the dynamic adjustment unit modifies the weight based on historical accident data and the operating environment, so that the weight can change with changes in the actual driving environment. The nonlinear interaction unit further introduces a coupling effect between events in the combined event processing. When multiple risk behaviors occur simultaneously, the risk contribution is enhanced and superimposed by introducing an amplification factor, thereby generating a combined event weight value that is more consistent with the actual risk level.

[0065] This weight allocation unit enables refined and intelligent allocation of risk weights. For example, a single sudden braking event might have a low risk weight on urban roads, but it would be dynamically adjusted to a higher value on mountain roads. When sudden acceleration and sharp turns are triggered simultaneously, the system not only considers their individual risk weights but also amplifies their combined effect through a nonlinear interaction unit, making the risk results more closely aligned with the patterns of actual accident occurrence. This approach enables the system to output a reasonable set of event risk characteristics in different scenarios, thereby improving the accuracy and adaptability of the overall rating.

[0066] In one embodiment of the present invention, the initial weight generation unit can search for the corresponding initial risk weight value from a preset risk weight table based on the event type of the event set. For example, the initial weight of a sudden acceleration event is set to 0.2, a sudden braking event to 0.3, and a sharp turn event to 0.25.

[0067] The dynamic adjustment unit can modify the initial weights based on historical accident data and the operating environment. This modification process multiplies the risk level of a particular event type by its frequency of occurrence in historical accident samples and normalizes the results across all event types so that the sum of the risk weights remains at 1. For example, if the accident probability of sudden braking events is significantly higher than that of sudden acceleration events in historical data, the system will increase the weight of sudden braking and decrease the weight of sudden acceleration.

[0068] The nonlinear interaction unit can introduce coupling effects between events in combined event processing. When two or more dynamic behavior features are triggered simultaneously within the same time window, the system multiplies the statistical value of each feature by the corresponding dynamically adjusted weight value to obtain a single risk contribution. These risk contributions are then amplified and superimposed using a preset amplification factor. For example, when sudden acceleration and sharp turns occur simultaneously, the system multiplies the original risk contribution by an amplification factor of 1.5, resulting in a higher combined event risk weight value that more realistically reflects the increased accident risk caused by the coupled driving behaviors.

[0069] Through the above settings, the weight allocation unit can achieve refined weighting of risk events, so that the system can not only reflect the risk of a single driving behavior, but also accurately capture the superimposed risk effects under complex driving behaviors, thereby improving the rationality and accuracy of the overall risk assessment results.

[0070] In a preferred embodiment of the present invention, the road condition assessment unit includes: A basic level retrieval unit is used to compare the location information of the target operating truck with a preset road condition risk level table to determine the initial road condition risk level; An external data acquisition unit is used to obtain external environmental data of the road section where the vehicle is located when an incident occurs, wherein the external environmental data includes the degree of road congestion, real-time weather conditions and historical accident frequency, and generate external road condition data; The level correction unit is used to correct the initial road condition risk level according to the external road condition data. When the external environment data indicates that the road risk increases, the risk level is increased. When the external environment data indicates that the road risk decreases, the risk level is reduced to obtain a dynamically updated road condition risk level.

[0071] In an embodiment of the present invention, dynamic risk correction of the vehicle operating environment is achieved by providing a basic level retrieval unit, an external data acquisition unit, and a level correction unit in the road condition assessment unit. The basic level retrieval unit can quickly match the initial level in a preset road condition risk level table based on the vehicle's location information, such as a low level for expressways and a high level for mountainous roads. The external data acquisition unit further introduces real-time road status data, including traffic congestion index, real-time weather information, and the historical accident frequency of the road section, thereby compensating for the limitations of relying solely on a fixed level table. The level correction unit corrects the initial level based on these real-time external data, so that the road condition risk level can be dynamically updated as environmental conditions change.

[0072] This road condition assessment unit ensures the timeliness and accuracy of risk assessment results. For example, the risk of a highway on a sunny day differs significantly from that on a foggy day. By incorporating external information such as weather and historical accident rates, the risk level can be adjusted from 1 to 3, making the event risk signature set more accurate to the actual operating environment. This not only avoids underestimation of risk but also provides a more refined basis for insurance pricing and scheduling management.

[0073] In one embodiment of the present invention, the basic level search unit can retrieve an initial road risk level from a preset road risk level table based on the target truck's location information. For example, expressways can be assigned level 1, urban roads level 2, mountainous roads level 3, and particularly dangerous sections (e.g., sharp bends and long downhill slopes) level 4.

[0074] The external data collection unit can capture environmental data about the road section where the vehicle is located when an incident occurs. This data can be collected through onboard cameras, traffic management platforms, or meteorological service interfaces. Specifically, it includes: road congestion index (e.g., levels 1-5), weather conditions (e.g., sunny, rainy, foggy, snowy, etc.), and historical accident frequency for that road section (e.g., number of accidents or accident rate over the past year).

[0075] The level correction unit can modify the initial road risk level based on external data. For example, if the system identifies a vehicle traveling on a rainy highway, it can raise the initial risk level from 1 to 2. If it detects a high accident rate on that highway section, it can raise the level to 3. Conversely, when the road is clear and the weather is good, the system can maintain or lower the risk level. The resulting dynamically updated road risk level more accurately reflects the actual risk level of the vehicle's environment.

[0076] In a preferred embodiment of the present invention, the weighted fusion unit includes: A coefficient optimization unit is used to iteratively optimize preset weighting coefficients by training multi-dimensional data samples of historical operating trucks to generate optimized weighting coefficients. The iterative optimization process includes: inputting historical samples into the machine learning model, comparing the risk feature vectors output by the model with the corresponding accident records, and gradually adjusting the weighting coefficients until convergence; The time series weighting unit is used to fuse the event risk feature set and historical static information parameters in the fusion input data according to the optimized weighting coefficient. In the processing, a time decay factor is introduced to reduce the weight of data at earlier moments and increase the weight of recent data to obtain the time series weighted result. The nonlinear fusion unit is used to perform nonlinear mapping and weighted summation on the interaction between the event risk feature set and historical static information based on the time series weighted results, and generate a multidimensional risk feature vector containing static factors and dynamic factors.

[0077] In this embodiment of the present invention, the intelligent fusion of dynamic and static information is achieved by incorporating a coefficient optimization unit, a time series weighting unit, and a nonlinear fusion unit within the weighted fusion unit. The coefficient optimization unit iteratively optimizes the preset weighting coefficients using historical operational truck data samples, ensuring that the fusion process continuously approximates the actual accident distribution and enhancing the model's predictive capabilities. The time series weighting unit introduces a time decay factor into the weighting process, giving higher weight to recent data and lower weight to long-term data, thereby highlighting the sensitive impact of short-term driving behavior on risk levels. The nonlinear fusion unit further nonlinearly maps and weights the interactions between the event risk feature set and historical static information to generate a risk feature vector containing multidimensional factors.

[0078] This weighted fusion unit significantly improves the accuracy and robustness of risk profile representation. For example, if a vehicle has frequently braked suddenly over the past week, even if it has a good annual inspection record, the system will amplify recent driving risks through time-series weighting and nonlinear fusion, making the risk rating more consistent with the current operating status. Compared to traditional linear weighting methods, this fusion method can more accurately characterize risk levels in complex driving scenarios.

[0079] In one embodiment of the present invention, the coefficient optimization unit iteratively optimizes preset weighting coefficients by training on multidimensional data samples of historical truck operations. For example, 100,000 vehicle operation data samples from a historical accident database are selected as training data and fed into a machine learning model (such as logistic regression, random forest, or neural network). The risk feature vectors output by the model are compared with actual accident records. The weighting coefficients are then gradually adjusted using a gradient descent algorithm until the error falls below a set threshold of 0.01 or the number of iterations reaches 500, resulting in the optimized weighting coefficients.

[0080] The time series weighting unit can introduce a time decay factor into the fusion process to emphasize the risk contribution of recent driving behavior. For example, an exponential decay function can be used as the weighting method. The closer the event to the current time, the higher the weight of the risk feature set; the farther away the event is, the lower the weight is. In this way, a sudden braking event within the past 24 hours will have a more significant impact on the risk assessment, while the same event from a week ago will have a smaller impact.

[0081] The nonlinear fusion unit further maps and weights the interaction between the event risk feature set and historical static information through nonlinear functions. For example, using a piecewise function, if a vehicle has a high number of annual inspection failures and has recently exhibited frequent rapid acceleration, the system will amplify the contribution of this combined feature to the risk vector. When both static and dynamic features are at normal levels, the system will suppress their influence on the risk vector. The resulting multidimensional risk feature vector more accurately reflects the vehicle's true risk level across different time periods and environments.

[0082] In a preferred embodiment of the present invention, the nonlinear fusion unit includes: The feature interaction modeling unit is used to model the feature interaction relationship between the event risk feature set and historical static information based on the time series weighted results, convert the interaction strength of the two types of features into interaction feature values, and form the interaction modeling results; A nonlinear mapping unit is used to perform nonlinear function transformation on the dynamic behavior characteristics, event risk weights, and road condition risk levels based on the interactive modeling results, and to combine the mapping values ​​in a layer-by-layer iterative manner to generate a nonlinear mapping result, wherein the nonlinear function is a piecewise function that can amplify the contribution of high-risk features and suppress the noise of low-risk features; The fusion generation unit is used to perform weighted summation of dynamic behavior characteristics and static information parameters according to the nonlinear mapping results to obtain a multidimensional risk feature vector containing feature interactions.

[0083] In an embodiment of the present invention, by setting a feature interaction modeling unit, a nonlinear mapping unit and a fusion generation unit in a nonlinear fusion unit, deep interactive modeling and enhanced representation of risk features are achieved. The feature interaction modeling unit can identify the interactive relationship between dynamic behavior features and static information, such as "vehicle age × sudden braking frequency", and convert it into a new interactive feature value. The nonlinear mapping unit performs nonlinear transformation on these interactive features through neural network stacking calculation or piecewise function mapping, so that high-risk combinations are amplified and low-risk combinations are suppressed, thereby improving the discrimination of feature expression. The fusion generation unit performs a weighted summation of the nonlinear mapping result and the original feature to obtain a multidimensional risk feature vector containing interactions.

[0084] This nonlinear fusion unit significantly enhances the ability to express risk features. For example, if two vehicles brake suddenly on the same road section, but one is over 10 years old and the other is only two years old, the system can highlight the higher risk of the former through interactive modeling and nonlinear mapping, thereby outputting a more realistic risk rating. This approach not only identifies potential risks that are difficult to reflect with a single feature, but also improves the accuracy and reliability of risk prediction in complex scenarios.

[0085] In one embodiment of the present invention, the feature interaction modeling unit can model the interaction between the event risk feature set and historical static information based on the time-series weighted results. For example, a vehicle's age and the number of sudden braking events can be combined into a cross-feature item to reflect the high risk of frequent sudden braking in older vehicles; vehicle maintenance records and abnormal fuel consumption can be combined into a cross-feature to reflect potential mechanical hazards.

[0086] The nonlinear mapping unit can map these cross-features using nonlinear functions, such as using a sigmoid function to amplify the impact of high-risk combinations or a piecewise linear function to suppress the interference of low-risk features. In a preferred embodiment, when a vehicle is over 10 years old and the frequency of sudden braking exceeds a threshold, the system uses a nonlinear function to increase the risk weight to twice the original value. For vehicles as young as two years old, even if sudden braking occurs, the weight increase is kept to a low level.

[0087] The fusion generation unit performs a weighted summation of the nonlinear mapping results with the original dynamic behavior features and static information to generate a multidimensional risk feature vector that incorporates feature interactions. For example, the final vector can contain 20 dimensions, including both single-event and road condition risks, as well as the "vehicle age x number of sudden braking" feature value derived through cross-modeling. This vector more comprehensively characterizes the risk characteristics of a vehicle in complex operating scenarios, thereby improving the accuracy and robustness of risk rating results.

[0088] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A multi-dimensional data fusion system for risk rating of operating trucks, characterized by: The system comprises: The data acquisition module is used to obtain the dynamic operation data and historical static information of the target operating trucks to form a preliminary multi-dimensional data set; The data preprocessing module is used to align fields, remove outliers, and normalize the format of the preliminary multidimensional data set to generate structured multidimensional data; The feature analysis module is used to extract the dynamic behavior characteristics of vehicle operation from structured multidimensional data and perform statistical calculations within a preset time window to form a feature statistical value data set; The event identification and risk assignment module is used to identify single events or combined events based on the characteristic statistical value data set and preset event trigger conditions, and assign event risk weights to them. At the same time, it combines the vehicle location information to determine the corresponding road condition risk level and summarizes it to form an event risk feature set; The multidimensional fusion module is used to fuse the event risk feature set with the historical static information in the structured multidimensional data to generate a time-series multidimensional risk feature vector; The risk rating module is used to score the multi-dimensional risk feature vector and output the corresponding risk rating results.

2. The multi-dimensional data fusion system for operating truck risk rating according to claim 1 is characterized in that: The feature analysis module includes: A dynamic feature segmentation unit is used to divide the dynamic operation data into multiple time periods according to a preset time window to obtain time segmentation data; The characteristic value calculation unit is used to process the time-segmented data, calculate the instantaneous value and cumulative value of the dynamic behavior characteristics, and form a characteristic value set; The statistics generation unit is used to calculate the change amplitude, peak value and cumulative impact of each dynamic behavior feature within a time period based on the feature value set, and generate a feature statistical value data set.

3. The multi-dimensional data fusion system for operating truck risk rating according to claim 1 is characterized in that: The event identification and risk assessment module includes: An event identification unit is used to identify a single event or a combination of events of a target operating truck within a preset time period based on a feature statistical value data set and preset trigger conditions to obtain an event set; The weight allocation unit is used to assign a preset risk weight value to a single event according to the event type of the event set, and perform a weighted operation on the statistical value of each dynamic behavior feature of the combined event and the corresponding risk weight to obtain the event risk weight value; The road condition assessment unit is used to combine the location information of the event set, search the preset road condition risk level table, determine the corresponding road condition risk level, and summarize it with the road condition risk level and event risk weight value to form an event risk feature set.

4. The multi-dimensional data fusion system for operating truck risk rating according to claim 1 is characterized in that: The multi-dimensional fusion module includes: A feature matching unit is used to associate and match the event risk feature set with the historical static information of the target operating truck based on the unique identifier of the target operating truck and the timestamp of the event, to obtain fused input data corresponding to the event; The weighted fusion unit is used to multiply the event risk feature set and the historical static information parameters by the preset weighting coefficients according to the fusion input data, and then sum them up to generate a multidimensional risk feature vector containing static factors and dynamic factors; The sequence generating unit is used to arrange the multi-dimensional risk feature vectors in chronological order to form a multi-dimensional risk feature vector sequence.

5. The multi-dimensional data fusion system for operating truck risk rating according to claim 2 is characterized in that: The statistics generating unit includes: An amplitude calculation unit, configured to calculate the difference between the maximum value and the minimum value of each dynamic behavior feature within a time period based on the feature value set, and generate an amplitude calculation result; The peak extraction unit is used to analyze the time series of dynamic behavior characteristics based on the amplitude calculation results, identify local maximum values ​​and record their size and occurrence time to generate peak extraction results; The cumulative statistical unit is used to accumulate all dynamic behavior feature values ​​within the time period item by item based on the peak extraction result, and weight the values ​​of different time segments with preset weights to form a cumulative impact statistical result; The result generating unit is used to combine the amplitude calculation result, the peak value extraction result and the cumulative impact quantity statistical result to generate a characteristic statistical value data set.

6. The multi-dimensional data fusion system for operating truck risk rating according to claim 3 is characterized in that: The event determination unit includes: a single event detection unit, configured to compare the statistical value of each dynamic behavior feature with the corresponding preset trigger condition based on the feature statistical value data set, and generate a single event record when the statistical value of a dynamic behavior feature exceeds a preset first number threshold; a combined event identification unit, configured to analyze multiple single events within the same time window based on the single event record, and combine the single events into a combined event when a preset combination condition is met, thereby generating a combined event record, wherein the preset combination condition is that different types of single events are triggered simultaneously within the same time window, or that the number of times the same type of single event is triggered exceeds a preset second number threshold within a preset time interval; The event set generating unit is configured to aggregate the single event record and the combined event record to generate an event set.

7. The multi-dimensional data fusion system for operating truck risk rating according to claim 3 is characterized in that: The weight distribution unit includes: An initial weight generating unit, configured to determine an initial risk weight value based on the event type of the event set and by comparing it with a preset risk weight table; A dynamic adjustment unit is used to modify the initial risk weight value based on historical accident data and operating environment data to generate a dynamically adjusted event risk weight value. The modification process is as follows: the risk level corresponding to different types of events is multiplied by their occurrence frequency in historical accident samples, and all product results are normalized; The nonlinear interaction unit is used to process the combined events according to the dynamically adjusted event risk weight value. When two or more dynamic behavior features are triggered simultaneously in the same time window, the statistical value of each feature is multiplied by the corresponding dynamically adjusted weight value to obtain a single risk contribution. Then, multiple risk contributions are enhanced and superimposed according to the preset amplification factor to generate a combined event weight value.

8. The multi-dimensional data fusion system for operating truck risk rating according to claim 3 is characterized in that: The road condition assessment unit includes: A basic level retrieval unit is used to compare the location information of the target operating truck with a preset road condition risk level table to determine the initial road condition risk level; An external data acquisition unit is used to obtain external environmental data of the road section where the vehicle is located when an incident occurs, wherein the external environmental data includes the degree of road congestion, real-time weather conditions and historical accident frequency, and generate external road condition data; The level correction unit is used to correct the initial road condition risk level according to the external road condition data. When the external environment data indicates that the road risk increases, the risk level is increased. When the external environment data indicates that the road risk decreases, the risk level is reduced to obtain a dynamically updated road condition risk level.

9. The multi-dimensional data fusion system for operating truck risk rating according to claim 4 is characterized in that: The weighted fusion unit includes: A coefficient optimization unit is used to iteratively optimize preset weighting coefficients by training multi-dimensional data samples of historical operating trucks to generate optimized weighting coefficients. The iterative optimization process includes: inputting historical samples into the machine learning model, comparing the risk feature vectors output by the model with the corresponding accident records, and gradually adjusting the weighting coefficients until convergence; The time series weighting unit is used to fuse the event risk feature set and historical static information parameters in the fusion input data according to the optimized weighting coefficient. In the processing, a time decay factor is introduced to reduce the weight of data at earlier moments and increase the weight of recent data to obtain the time series weighted result. The nonlinear fusion unit is used to perform nonlinear mapping and weighted summation on the interaction between the event risk feature set and historical static information based on the time series weighted results, and generate a multidimensional risk feature vector containing static factors and dynamic factors.

10. The multi-dimensional data fusion system for operating truck risk rating according to claim 9 is characterized in that: The nonlinear fusion unit includes: The feature interaction modeling unit is used to model the feature interaction relationship between the event risk feature set and historical static information based on the time series weighted results, convert the interaction strength of the two types of features into interaction feature values, and form the interaction modeling results; A nonlinear mapping unit is used to perform nonlinear function transformation on the dynamic behavior characteristics, event risk weights, and road condition risk levels based on the interactive modeling results, and to combine the mapping values ​​in a layer-by-layer iterative manner to generate a nonlinear mapping result, wherein the nonlinear function is a piecewise function that can amplify the contribution of high-risk features and suppress the noise of low-risk features; The fusion generation unit is used to perform weighted summation of dynamic behavior characteristics and static information parameters according to the nonlinear mapping results to obtain a multidimensional risk feature vector containing feature interactions.

Images