T-box data-driven vehicle driving risk dynamic evaluation method and system

Abstract

The application discloses a kind of T-Box data-driven vehicle driving risk dynamic evaluation method and system, the method includes: obtaining the multiple communication data uploaded by the T-Box equipment of the vehicle to be evaluated;Based on data analysis algorithm, determine the driving characteristic and environmental characteristic corresponding to each communication data;According to the driving characteristic and environmental characteristic corresponding to all the communication data, based on recurrent neural network, predict future driving behavior;According to the future driving behavior, and risk accumulation calculation rule, calculate the driving risk parameter corresponding to the vehicle to be identified.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a T-Box data-driven method and system for dynamic assessment of vehicle driving risks. Background Technology

[0002] With the rapid development of intelligent connected vehicles and vehicle safety monitoring, automotive companies and regulatory platforms are increasingly emphasizing the monitoring of communication data from vehicle T-Box devices to assist in assessing driving risks. Improving the accuracy of risk assessment has become a key technical issue. Existing technologies typically acquire communication data uploaded by the vehicle's T-Box device, employing fixed threshold analysis or simple statistical methods to evaluate current driving behavior and making risk judgments based on historical data to support vehicle safety management. However, existing solutions lack dynamic correlation analysis of driving and environmental characteristics, the application of recurrent neural networks for predicting future driving behavior, and the use of risk accumulation calculation rules. This makes it difficult to achieve in-depth temporal analysis of driving behavior and forward-looking risk prediction. Commonly used static or isolated historical data analysis strategies fail to capture changes in driving trends, resulting in insufficient foresight and accuracy in driving risk assessment. The lack of correlation between historical data and future trends can easily lead to the omission of safety hazards, limiting the intelligence level and actual protective effect of vehicle driving risk monitoring. Therefore, existing technologies have shortcomings that urgently need to be addressed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a T-Box data-driven method and system for dynamic assessment of vehicle driving risks, which can realize the temporal feature analysis of vehicle driving behavior and intelligent prediction of future risks driven by T-Box data, improve the foresight and accuracy of driving risk assessment, and reduce the risk of missing safety hazards due to historical data not being associated with future trends.

[0004] To address the aforementioned technical problems, the first aspect of this invention discloses a T-Box data-driven method for dynamic assessment of vehicle driving risks, the method comprising: Acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated; Based on data analysis algorithms, the driving characteristics and environmental characteristics corresponding to each piece of communication data are determined. Based on the driving and environmental characteristics corresponding to all the aforementioned communication data, future driving behavior is predicted using a recurrent neural network. Based on the future driving behavior and the risk accumulation calculation rules, the driving risk parameters corresponding to the vehicle to be identified are calculated.

[0005] As an optional implementation, in the first aspect of the invention, determining the driving characteristics and environmental characteristics corresponding to each piece of communication data based on a data analysis algorithm includes: For each of the aforementioned communication data, based on character recognition rules, the vehicle driving data, environmental sensing data, and vehicle system operating data in the communication data are identified; Based on the vehicle driving data, analyze the driving characteristics corresponding to the communication data; Based on the environmental sensing data and the vehicle system operating data, the environmental characteristics corresponding to the communication data are analyzed.

[0006] As an optional implementation, in the first aspect of the present invention, the step of analyzing the driving characteristics corresponding to the communication data based on the vehicle driving data includes: Based on the vehicle driving data, determine the vehicle acceleration, vehicle direction of travel, and vehicle driving operation of the vehicle to be evaluated corresponding to the communication data; Based on the preset threshold rules for the three urgent behaviors, as well as the vehicle acceleration and the vehicle's direction of travel, it is determined whether the vehicle to be evaluated exhibits the three urgent behaviors, and a first judgment result is obtained. When the first determination result is yes, the vehicle driving operation, the vehicle acceleration, and the corresponding three emergency behavior types are determined as the driving characteristics corresponding to the communication data; the three emergency behavior types are rapid acceleration, sharp turning, or rapid deceleration. If the first determination result is negative, the vehicle driving operation is determined as the driving feature corresponding to the communication data.

[0007] As an optional implementation, in the first aspect of the present invention, the step of analyzing the environmental characteristics corresponding to the communication data based on the environmental sensing data and the vehicle-mounted system operating data includes: The environmental sensing data is input into the trained in-vehicle activity prediction model to obtain the output predicted in-vehicle behavior; the in-vehicle activity prediction model is trained using a training dataset that includes multiple training environmental sensing data and corresponding in-vehicle behavior annotations. Determine whether the predicted in-vehicle behavior includes any preset violations to obtain a second determination result; the violations include at least one of the following: not wearing a seat belt, having children but not using a child seat, the driver making a phone call, overloading the vehicle, and improperly placing the seats in the vehicle. When the second judgment result is yes, based on the matching model corresponding to the existing violation, the environmental characteristics corresponding to the communication data are determined according to the vehicle system working data; If the second determination result is negative, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0008] As an optional implementation, in the first aspect of the present invention, determining the environmental characteristics corresponding to the communication data based on the vehicle-mounted system's operating data using a matching model corresponding to existing violations includes: The matching prediction model corresponding to the existing violation is determined in the preset matching model database; the matching prediction model is trained by a training dataset including multiple vehicle system working reference data and corresponding annotations indicating whether the violation exists; the vehicle system working reference data includes at least one of vehicle system seat belt warning data, vehicle system seat perception data, vehicle system call working data and vehicle system control operation data. The vehicle system's operating data is input into the matching prediction model to obtain the output matching probability; When the matching probability is greater than a preset probability threshold, the existing violation and the corresponding matching probability are determined as the environmental characteristics corresponding to the communication data. When the matching probability is less than the probability threshold, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0009] As an optional implementation, in the first aspect of the invention, the step of predicting future driving behavior based on a recurrent neural network according to the driving characteristics and environmental characteristics corresponding to all the communication data includes: The data sequence is obtained by sorting all the communication data from morning to night based on the corresponding upload time points; The corresponding driving features and environmental features are determined as data feature labels for each communication data in the data sequence to obtain an updated labeled data sequence; Based on a recurrent neural network, future driving behavior is predicted according to the labeled data sequence.

[0010] As an optional implementation, in the first aspect of the invention, predicting future driving behavior based on the labeled data sequence using a recurrent neural network includes: The labeled data sequence is input into a trained recurrent neural network model to obtain the output future driving behavior; the recurrent neural network model is trained using a training dataset that includes driving behavior sequences of multiple training vehicles and corresponding T-Box device communication data annotations, driving feature annotations, and environmental feature annotations; the future driving behavior includes multiple driving operations ordered according to their possible time.

[0011] As an optional implementation, in the first aspect of the present invention, calculating the driving risk parameters corresponding to the vehicle to be identified based on the future driving behavior and risk accumulation calculation rules includes: Determine the current risk parameters corresponding to the vehicle to be identified; the current risk parameters are calculated based on the driving characteristics corresponding to the communication data using a preset risk scoring rule; Based on the risk scoring rules and the future driving behavior, calculate the cumulative risk parameters; Calculate the similarity between the future driving behavior and the current driving feature set, and calculate the parameter weights that are proportional to the similarity; the current driving feature set includes multiple driving features. The modified cumulative risk parameter is obtained by multiplying the parameter weights and the cumulative risk parameter. The sum of the current risk parameter and the corrected cumulative risk parameter is calculated to obtain the driving risk parameter corresponding to the vehicle to be identified.

[0012] A second aspect of this invention discloses a T-Box data-driven dynamic assessment system for vehicle driving risks, the system comprising: The acquisition module is used to acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated; The determination module is used to determine the driving characteristics and environmental characteristics corresponding to each piece of communication data based on data analysis algorithms; The prediction module is used to predict future driving behavior based on a recurrent neural network, according to the driving characteristics and environmental characteristics corresponding to all the communication data. The calculation module is used to calculate the driving risk parameters corresponding to the vehicle to be identified based on the future driving behavior and the risk accumulation calculation rules.

[0013] As an optional implementation, in a second aspect of the invention, the determining module determines the specific method by which it determines the driving characteristics and environmental characteristics corresponding to each piece of communication data based on a data analysis algorithm, including: For each of the aforementioned communication data, based on character recognition rules, the vehicle driving data, environmental sensing data, and vehicle system operating data in the communication data are identified; Based on the vehicle driving data, analyze the driving characteristics corresponding to the communication data; Based on the environmental sensing data and the vehicle system operating data, the environmental characteristics corresponding to the communication data are analyzed.

[0014] As an optional implementation, in a second aspect of the invention, the specific method by which the determining module analyzes the driving characteristics corresponding to the communication data based on the vehicle driving data includes: Based on the vehicle driving data, determine the vehicle acceleration, vehicle direction of travel, and vehicle driving operation of the vehicle to be evaluated corresponding to the communication data; Based on the preset threshold rules for the three urgent behaviors, as well as the vehicle acceleration and the vehicle's direction of travel, it is determined whether the vehicle to be evaluated exhibits the three urgent behaviors, and a first judgment result is obtained. When the first determination result is yes, the vehicle driving operation, the vehicle acceleration, and the corresponding three emergency behavior types are determined as the driving characteristics corresponding to the communication data; the three emergency behavior types are rapid acceleration, sharp turning, or rapid deceleration. If the first determination result is negative, the vehicle driving operation is determined as the driving feature corresponding to the communication data.

[0015] As an optional implementation, in a second aspect of the invention, the specific method by which the determining module analyzes the environmental characteristics corresponding to the communication data based on the environmental sensing data and the vehicle-mounted system operating data includes: The environmental sensing data is input into the trained in-vehicle activity prediction model to obtain the output predicted in-vehicle behavior; the in-vehicle activity prediction model is trained using a training dataset that includes multiple training environmental sensing data and corresponding in-vehicle behavior annotations. Determine whether the predicted in-vehicle behavior includes any preset violations to obtain a second determination result; the violations include at least one of the following: not wearing a seat belt, having children but not using a child seat, the driver making a phone call, overloading the vehicle, and improperly placing the seats in the vehicle. When the second judgment result is yes, based on the matching model corresponding to the existing violation, the environmental characteristics corresponding to the communication data are determined according to the vehicle system working data; If the second determination result is negative, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0016] As an optional implementation, in a second aspect of the invention, the determining module determines the specific method by which it determines the environmental characteristics corresponding to the communication data based on the vehicle-mounted system's operating data, according to a matching model corresponding to existing violations, including: The matching prediction model corresponding to the existing violation is determined in the preset matching model database; the matching prediction model is trained by a training dataset including multiple vehicle system working reference data and corresponding annotations indicating whether the violation exists; the vehicle system working reference data includes at least one of vehicle system seat belt warning data, vehicle system seat perception data, vehicle system call working data and vehicle system control operation data. The vehicle system's operating data is input into the matching prediction model to obtain the output matching probability; When the matching probability is greater than a preset probability threshold, the existing violation and the corresponding matching probability are determined as the environmental characteristics corresponding to the communication data. When the matching probability is less than the probability threshold, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0017] As an optional implementation, in a second aspect of the invention, the prediction module predicts the specific manner of future driving behavior based on a recurrent neural network, according to the driving characteristics and environmental characteristics corresponding to all the communication data, including: The data sequence is obtained by sorting all the communication data from morning to night based on the corresponding upload time points; The corresponding driving features and environmental features are determined as data feature labels for each communication data in the data sequence to obtain an updated labeled data sequence; Based on a recurrent neural network, future driving behavior is predicted according to the labeled data sequence.

[0018] As an optional implementation, in a second aspect of the invention, the prediction module is based on a recurrent neural network and predicts the specific manner of future driving behavior according to the labeled data sequence, including: The labeled data sequence is input into a trained recurrent neural network model to obtain the output future driving behavior; the recurrent neural network model is trained using a training dataset that includes driving behavior sequences of multiple training vehicles and corresponding T-Box device communication data annotations, driving feature annotations, and environmental feature annotations; the future driving behavior includes multiple driving operations ordered according to their possible time.

[0019] As an optional implementation, in a second aspect of the invention, the specific method by which the calculation module calculates the driving risk parameters corresponding to the vehicle to be identified based on the future driving behavior and risk accumulation calculation rules includes: Determine the current risk parameters corresponding to the vehicle to be identified; the current risk parameters are calculated based on the driving characteristics corresponding to the communication data using a preset risk scoring rule; Based on the risk scoring rules and the future driving behavior, calculate the cumulative risk parameters; Calculate the similarity between the future driving behavior and the current driving feature set, and calculate the parameter weights that are proportional to the similarity; the current driving feature set includes multiple driving features. The modified cumulative risk parameter is obtained by multiplying the parameter weights and the cumulative risk parameter. The sum of the current risk parameter and the corrected cumulative risk parameter is calculated to obtain the driving risk parameter corresponding to the vehicle to be identified.

[0020] A third aspect of this invention discloses another T-Box data-driven dynamic assessment system for vehicle driving risks, the system comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the T-Box data-driven dynamic assessment method for vehicle driving risks disclosed in the first aspect of the present invention.

[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the T-Box data-driven dynamic assessment method for vehicle driving risks disclosed in the first aspect of the present invention.

[0022] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention acquires multiple communication data uploaded by the T-Box device of the vehicle to be evaluated, determines driving characteristics and environmental characteristics based on data analysis algorithms, predicts future driving behavior based on both using a recurrent neural network, and calculates driving risk parameters based on future driving behavior and risk accumulation calculation rules. This enables T-Box data-driven temporal feature analysis of vehicle driving behavior and intelligent prediction of future risks, improving the foresight and accuracy of driving risk assessment and reducing the risk of missing safety hazards due to historical data not being correlated with future trends. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a T-Box data-driven dynamic assessment method for vehicle driving risks disclosed in an embodiment of the present invention.

[0025] Figure 2 This is a schematic diagram of the structure of a T-Box data-driven dynamic assessment system for vehicle driving risks disclosed in an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram of another T-Box data-driven dynamic assessment system for vehicle driving risks disclosed in an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] This invention discloses a T-Box data-driven method and system for dynamic assessment of vehicle driving risks. By acquiring multiple communication data uploaded from the T-Box device of the vehicle to be assessed, driving characteristics and environmental characteristics are determined based on data analysis algorithms. Future driving behavior is predicted using a recurrent neural network based on both characteristics. Driving risk parameters are calculated based on future driving behavior and risk accumulation calculation rules. This enables time-series feature analysis of vehicle driving behavior and intelligent prediction of future risks driven by T-Box data, improving the foresight and accuracy of driving risk assessment and reducing the risk of overlooking safety hazards due to historical data not being correlated with future trends. Detailed explanations follow.

[0031] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating a T-Box data-driven dynamic assessment method for vehicle driving risk disclosed in an embodiment of the present invention. Figure 1 The described T-Box data-driven dynamic assessment method for vehicle driving risks can be applied to data processing systems / data processing devices / data processing servers (wherein, the server includes a local processing server or a cloud processing server). For example... Figure 1As shown, the T-Box data-driven dynamic assessment method for vehicle driving risk may include the following operations: 101. Obtain multiple communication data uploaded by the T-Box device of the vehicle to be evaluated.

[0032] Optionally, the vehicle to be evaluated can be a heavy-duty diesel truck of a logistics fleet, a passenger car permanently stationed on a ride-hailing platform, or a special tank tractor for transporting hazardous chemicals; this invention does not impose any limitations.

[0033] Optionally, the T-Box device can be an automotive-grade in-vehicle intelligent terminal that integrates a five-mode full-network cellular communication module, a high-precision GNSS positioning chip, and a dual-channel high-speed CAN bus controller; this invention does not limit the scope of the device.

[0034] Optionally, the communication data can be a custom binary stream message reported at high frequency based on the MQTT protocol, a lightweight data packet that has been structured and serialized using Protocol Buffers, or an encrypted time-series message that conforms to the national standard GB / T 32960. This invention does not impose any limitations on this.

[0035] 102. Based on data analysis algorithms, determine the driving characteristics and environmental characteristics corresponding to each communication data. Optionally, the data analysis algorithm can be a real-time rule filtering algorithm based on the streaming computing engine Flink, feature extraction rules based on spatiotemporal correlation graphs, or unstructured data parsing logic based on sliding windows; this invention does not limit the specific algorithm.

[0036] 103. Based on the driving and environmental characteristics corresponding to all communication data, predict future driving behavior using a recurrent neural network. 104. Based on future driving behavior and risk accumulation calculation rules, calculate the driving risk parameters corresponding to the vehicle to be identified.

[0037] Optionally, the recurrent neural network can be a Long Short-Term Memory (LSTM) network with a gating mechanism, a lightweight gated recurrent unit (GRU), or a bidirectional recurrent neural network (BiRNN), and this invention does not limit it.

[0038] Optionally, the risk accumulation calculation rule can be a sliding window integral algorithm based on time exponential weight decay, a state transition accumulation rule based on Markov decision chain, or a risk scoring matrix iterative operator with a forgetting factor. This invention does not limit the specific rules.

[0039] Optionally, the driving risk parameter can be a quantitative safety coefficient used for differentiated pricing (UBI) in commercial auto insurance, a conditional probability score characterizing the vehicle's collision within a very short period in the future, or a high-risk behavior rating indicator in a commercial fleet management system. This invention does not limit the specific parameters.

[0040] As can be seen, the above-described embodiments of the invention acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated, determine driving characteristics and environmental characteristics based on data analysis algorithms, predict future driving behavior based on recurrent neural networks based on both, and calculate driving risk parameters based on future driving behavior and risk accumulation calculation rules. This enables the realization of time-series feature analysis and intelligent prediction of future risks of vehicle driving behavior driven by T-Box data, improves the foresight and accuracy of driving risk assessment, and reduces the risk of missing safety hazards due to historical data not being associated with future trends.

[0041] As an optional embodiment, the step described above, determining the driving characteristics and environmental characteristics corresponding to each piece of communication data based on a data analysis algorithm, includes: For each piece of communication data, based on character recognition rules, the vehicle driving data, environmental sensor data, and vehicle system operating data in the communication data are identified; Based on vehicle driving data, analyze the driving characteristics corresponding to the communication data; Based on environmental sensor data and vehicle system operating data, the environmental characteristics corresponding to this communication data are analyzed.

[0042] Optionally, the character recognition rule can be a dynamic byte stream truncation rule for a specific hexadecimal header, a string pattern matching rule based on regular expressions, or a field strong type mapping rule based on JSON Schema definition; this invention does not limit the specific rules.

[0043] Optionally, the vehicle driving data can be the absolute steering wheel opening angle, accelerator pedal depth percentage, electronic parking brake pump pressure, and the activation status of the anti-lock braking system (ABS) that can be read directly via the CAN bus. This invention does not limit these parameters.

[0044] Optionally, the environmental sensing data can be the image grayscale stream output by the in-vehicle infrared advanced driver assistance system (DMS) camera, the audio Hertz signal collected by the in-vehicle panoramic microphone array, or the analog electrical signal output by the rain and light combination sensor installed on the windshield. This invention does not limit the data.

[0045] Optionally, the vehicle's operating data may be a snapshot of the operation log of the in-vehicle multimedia infotainment system (IVI), the on / off level signal of the seat belt buckle sensor circuit, or the connection and pairing status data of the in-vehicle Bluetooth module; this invention does not limit the data.

[0046] As can be seen, through the above optional embodiments, by identifying vehicle driving data, environmental sensing data and vehicle system working data based on character recognition rules for each communication data, and analyzing driving characteristics and environmental characteristics respectively, the multi-type structured parsing and feature extraction of communication data can be realized, thereby improving the standardization and comprehensiveness of feature analysis and reducing the risk of feature extraction deviation caused by data mixing.

[0047] As an optional embodiment, the step above, analyzing the driving characteristics corresponding to the communication data based on vehicle driving data, includes: Based on the vehicle driving data, determine the vehicle acceleration, vehicle direction of travel, and vehicle driving operation of the vehicle to be evaluated corresponding to the communication data; Based on the preset threshold rules for the three urgent behaviors, as well as the vehicle acceleration and vehicle direction of travel, it is determined whether the vehicle to be evaluated exhibits the three urgent behaviors, and the first judgment result is obtained. When the first judgment result is yes, the vehicle driving operation, vehicle acceleration, and the corresponding three emergency behavior types are determined as the driving characteristics corresponding to the communication data; optionally, the three emergency behavior types are rapid acceleration, sharp turning, or rapid deceleration. If the first judgment result is negative, the vehicle driving operation is identified as the driving feature corresponding to the communication data.

[0048] Optionally, the vehicle acceleration can be a longitudinal linear acceleration scalar calculated by the six-axis inertial measurement unit (IMU) built into the vehicle body, and a lateral centrifugal acceleration scalar; this invention does not limit the specific acceleration.

[0049] Optionally, the vehicle's travel direction can be the absolute heading angle calculated based on the phase difference of dual-frequency GPS carriers, or the yaw rate variation calculated by combining the wheel speed difference; this invention does not limit the specific direction.

[0050] Optionally, the threshold rule for the three rapid behaviors can be set to limit the positive abrupt change in longitudinal gravitational acceleration to greater than 3.2 m / s². 2 Rapid acceleration threshold and negative braking speed greater than 4.5 m / s 2 The invention does not limit the rapid deceleration threshold or the adaptive curve equation for sharp turns where the product of lateral acceleration and vehicle speed exceeds the physical limit of sideslip.

[0051] As can be seen, through the above optional embodiments, by determining the vehicle acceleration, driving direction and driving operation based on vehicle driving data, judging whether there is an emergency behavior based on the emergency behavior threshold rule and determining the driving characteristics, the accurate identification of emergency behavior based on acceleration and operation rules is achieved, the dynamics and reliability of driving characteristics are improved, and the risk of misjudgment of behavior caused by fixed thresholds is reduced.

[0052] As an optional embodiment, the step above, analyzing the environmental characteristics corresponding to the communication data based on environmental sensing data and vehicle system operating data, includes: Environmental sensor data is input into a trained in-vehicle activity prediction model to obtain the output predicted in-vehicle behavior; optionally, the in-vehicle activity prediction model is trained using a training dataset that includes multiple training environmental sensor data and corresponding in-vehicle behavior annotations. Determine whether the predicted in-vehicle behavior constitutes a pre-set violation, and obtain a second judgment result; When the second judgment result is yes, based on the matching model corresponding to the existing violation, the environmental characteristics corresponding to the communication data are determined according to the vehicle system working data; If the second judgment result is negative, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0053] Optionally, violations include at least one of the following: not wearing a seatbelt, having children but not using a child seat, the driver using a mobile phone, overloading the vehicle, and improperly placing the seats in the vehicle.

[0054] Optionally, the in-vehicle activity prediction model can be a 3D human skeleton key point detection model based on the lightweight neural network MobileNet, a video behavior recognition network SlowFast that introduces a spatiotemporal attention mechanism, or a YOLOv8-pose human pose estimator deployed at the edge. This invention does not limit the model.

[0055] Optionally, the illegal placement of seats in the vehicle can be an abnormal form of unauthorized removal of rear seats from commercial vehicles for illegal cargo transport, or a dangerous placement state in which passenger car seats are folded down, causing the safety restraint system to fail. This invention does not limit this.

[0056] As can be seen, through the above optional embodiments, by inputting environmental sensing data into the in-vehicle activity prediction model to obtain predicted in-vehicle behavior, determining whether there is any violation, and determining environmental characteristics based on the matching model, intelligent analysis of environmental characteristics based on in-vehicle behavior prediction and violation matching is realized, thereby improving the violation sensitivity and accuracy of environmental characteristics and reducing the risk of insufficient environmental assessment due to failure to consider in-vehicle violations.

[0057] As an optional embodiment, the step described above, determining the environmental characteristics corresponding to the communication data based on the matching model corresponding to the existing violation and the vehicle-mounted system's operating data, includes: The matching prediction model corresponding to the existing violations is determined in the preset matching model database; the vehicle system's working data is input into the matching prediction model to obtain the output matching probability; When the matching probability is greater than a preset probability threshold, the existing violations and their corresponding matching probabilities are identified as environmental features corresponding to the communication data. When the matching probability is less than the probability threshold, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0058] Optionally, the matching prediction model is trained using a training dataset that includes multiple vehicle system operating reference data and corresponding annotations indicating whether violations exist.

[0059] Optionally, the vehicle infotainment system operating reference data may include at least one of the following: vehicle infotainment system seatbelt warning data, vehicle infotainment system seat perception data, vehicle infotainment system call operation data, and vehicle infotainment system control operation data.

[0060] Optionally, the matching model database can be a lightweight model weight dictionary embedded in the vehicle edge computing chip, or a high-speed key-value pair cache space deployed on a cloud server; this invention does not limit the scope of the database.

[0061] Optionally, the matching prediction model can be a binary classifier based on support vector machine (SVM), an extreme gradient boosting tree (XGBoost) classification model, or a deep multilayer perceptron (MLP) classification network; this invention does not limit the model.

[0062] Optionally, the matching probability can be a normalized risk probability value calculated by the softmax layer or sigmoid activation function at the end of the classification model; this invention does not limit this.

[0063] As can be seen, through the above optional embodiments, by determining the matching prediction model corresponding to the violation in the matching model database, inputting the vehicle system working data to obtain the matching probability, and determining the environmental characteristics based on the probability, the system can accurately quantify the violation environmental characteristics based on vehicle system data and model prediction, improve the intelligence and objectivity of environmental feature recognition, and reduce the risk of violation omissions due to single rules.

[0064] As an optional embodiment, the step described above, predicting future driving behavior based on a recurrent neural network according to the driving characteristics and environmental characteristics corresponding to all communication data, includes: The data sequence is obtained by sorting all communication data from morning to night based on the corresponding upload time points; The corresponding driving features and environmental features are determined as data feature labels for each communication data in the data sequence to obtain the updated labeled data sequence; Based on recurrent neural networks, future driving behavior can be predicted from labeled data sequences.

[0065] Optionally, the updated labeled data sequence can be a multimodal joint tensor time series sequence that is interwoven along the time axis and incorporates high-dimensional driving operators and sudden environmental factors; this invention does not impose any limitations on this.

[0066] As can be seen, through the above optional embodiments, by sorting communication data based on upload time points to obtain a data sequence, and labeling driving features and environmental features as data features to obtain a labeled data sequence, and predicting future driving behavior based on recurrent neural networks, the temporal labeling of historical communication data and the prediction of future behavior are realized, thereby improving the continuity and foresight of the prediction process and reducing the risk of prediction deviation caused by unlabeled time series.

[0067] As an optional embodiment, the step above, predicting future driving behavior based on a recurrent neural network and a labeled data sequence, includes: The labeled data sequence is input into a trained recurrent neural network model to obtain the output of future driving behavior.

[0068] Optionally, the recurrent neural network model is trained using a training dataset that includes driving behavior sequences of multiple training vehicles and corresponding T-Box device communication data annotations, driving feature annotations, and environmental feature annotations.

[0069] Optionally, future driving behavior includes multiple driving actions ordered in order of their possible timing.

[0070] Optionally, the future driving behavior may be an emergency lane change and steering operation, high-frequency braking operation, or continuous high-throttle overtaking control command that the driver may continuously perform in the next 3 seconds, 5 seconds, or 10 seconds. This invention does not limit the scope of the invention.

[0071] As can be seen, through the above optional embodiments, future driving behavior can be obtained by inputting labeled data sequences into a trained recurrent neural network model, thereby achieving deep prediction based on multi-labeled time series data, improving the accuracy and generalization ability of future driving behavior prediction, and reducing the risk of prediction errors caused by incomplete model input.

[0072] As an optional embodiment, the step above, calculating the driving risk parameters corresponding to the vehicle to be identified based on future driving behavior and risk accumulation calculation rules, includes: Determine the current risk parameters corresponding to the vehicle to be identified; optionally, the current risk parameters are calculated based on the driving characteristics corresponding to the communication data using preset risk scoring rules; Based on the risk scoring rules and future driving behavior, calculate the cumulative risk parameters; Calculate the similarity between future driving behavior and the current driving feature set, and calculate the parameter weights proportional to the similarity; optionally, the current driving feature set includes multiple driving features. The modified cumulative risk parameter is obtained by multiplying the parameter weights and the cumulative risk parameter. The sum of the current risk parameters and the corrected cumulative risk parameters is calculated to obtain the driving risk parameters corresponding to the vehicle to be identified.

[0073] Optionally, the similarity can be the distance similarity score between the future predicted trajectory sequence and the historical high-frequency driving behavior sequence calculated using the Dynamic Time Warping (DTW) algorithm, or the cosine similarity between two sets of temporal feature embedding vectors. This invention does not limit the similarity.

[0074] As can be seen, through the above optional embodiments, by determining the current risk parameters, calculating the cumulative risk parameters according to the risk scoring rules and future driving behavior, and calculating the weight to correct the cumulative risk parameters based on the similarity between future behavior and current driving characteristics, the driving risk parameters are finally obtained by calculating the sum of the current risk and the corrected cumulative risk. This achieves a fusion and quantitative assessment of current and future risks, improves the comprehensiveness and dynamism of driving risk parameters, and reduces the risk of underestimation due to failure to consider future behavior.

[0075] Example 2 Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of a T-Box data-driven dynamic assessment system for vehicle driving risks disclosed in an embodiment of the present invention. Figure 2 The described T-Box data-driven dynamic assessment system for vehicle driving risks can be applied to data processing systems / data processing devices / data processing servers (wherein, the server includes a local processing server or a cloud processing server). For example... Figure 2 As shown, the T-Box data-driven vehicle driving risk dynamic assessment system may include: The acquisition module 201 is used to acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated.

[0076] The determination module 202 is used to determine the driving characteristics and environmental characteristics corresponding to each piece of communication data based on data analysis algorithms. The prediction module 203 is used to predict future driving behavior based on a recurrent neural network, according to the driving characteristics and environmental characteristics corresponding to all communication data. The calculation module 204 is used to calculate the driving risk parameters corresponding to the vehicle to be identified based on future driving behavior and risk accumulation calculation rules.

[0077] As can be seen, the above-described embodiments of the invention acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated, determine driving characteristics and environmental characteristics based on data analysis algorithms, predict future driving behavior based on recurrent neural networks based on both, and calculate driving risk parameters based on future driving behavior and risk accumulation calculation rules. This enables the realization of time-series feature analysis and intelligent prediction of future risks of vehicle driving behavior driven by T-Box data, improves the foresight and accuracy of driving risk assessment, and reduces the risk of missing safety hazards due to historical data not being associated with future trends.

[0078] As an optional embodiment, the determining module determines the specific method by which it determines the driving characteristics and environmental characteristics corresponding to each piece of communication data based on a data analysis algorithm, including: For each piece of communication data, based on character recognition rules, the vehicle driving data, environmental sensor data, and vehicle system operating data in the communication data are identified; Based on vehicle driving data, analyze the driving characteristics corresponding to the communication data; Based on environmental sensor data and vehicle system operating data, the environmental characteristics corresponding to this communication data are analyzed.

[0079] As can be seen, through the above optional embodiments, by identifying vehicle driving data, environmental sensing data and vehicle system working data based on character recognition rules for each communication data, and analyzing driving characteristics and environmental characteristics respectively, the multi-type structured parsing and feature extraction of communication data can be realized, thereby improving the standardization and comprehensiveness of feature analysis and reducing the risk of feature extraction deviation caused by data mixing.

[0080] As an optional embodiment, the specific method by which the determining module analyzes the driving characteristics corresponding to the communication data based on vehicle driving data includes: Based on the vehicle driving data, determine the vehicle acceleration, vehicle direction of travel, and vehicle driving operation of the vehicle to be evaluated corresponding to the communication data; Based on the preset threshold rules for the three urgent behaviors, as well as the vehicle acceleration and vehicle direction of travel, it is determined whether the vehicle to be evaluated exhibits the three urgent behaviors, and the first judgment result is obtained. When the first judgment result is yes, the vehicle driving operation, vehicle acceleration, and the corresponding three emergency behavior types are determined as the driving characteristics corresponding to the communication data; optionally, the three emergency behavior types are rapid acceleration, sharp turning, or rapid deceleration. If the first judgment result is negative, the vehicle driving operation is identified as the driving feature corresponding to the communication data.

[0081] As can be seen, through the above optional embodiments, by determining the vehicle acceleration, driving direction and driving operation based on vehicle driving data, judging whether there is an emergency behavior based on the emergency behavior threshold rule and determining the driving characteristics, the accurate identification of emergency behavior based on acceleration and operation rules is achieved, the dynamics and reliability of driving characteristics are improved, and the risk of misjudgment of behavior caused by fixed thresholds is reduced.

[0082] As an optional embodiment, the specific method by which the determining module analyzes the environmental characteristics corresponding to the communication data based on environmental sensing data and vehicle system operating data includes: Environmental sensor data is input into a trained in-vehicle activity prediction model to obtain the output predicted in-vehicle behavior; optionally, the in-vehicle activity prediction model is trained using a training dataset that includes multiple training environmental sensor data and corresponding in-vehicle behavior annotations. Determine whether there are any pre-set violations in the predicted in-vehicle behavior to obtain a second judgment result; optionally, the violations include at least one of the following: not wearing a seat belt, having children but not using a child seat, the driver making a phone call, overloading the vehicle, and improper placement of seats in the vehicle; When the second judgment result is yes, based on the matching model corresponding to the existing violation, the environmental characteristics corresponding to the communication data are determined according to the vehicle system working data; If the second judgment result is negative, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0083] As can be seen, through the above optional embodiments, by inputting environmental sensing data into the in-vehicle activity prediction model to obtain predicted in-vehicle behavior, determining whether there is any violation, and determining environmental characteristics based on the matching model, intelligent analysis of environmental characteristics based on in-vehicle behavior prediction and violation matching is realized, thereby improving the violation sensitivity and accuracy of environmental characteristics and reducing the risk of insufficient environmental assessment due to failure to consider in-vehicle violations.

[0084] As an optional embodiment, the determining module determines the specific method by which it determines the environmental characteristics corresponding to the communication data based on the matching model corresponding to the existing violation and the vehicle-mounted system's operating data, including: The matching prediction model corresponding to the existing violation is determined in the preset matching model database; optionally, the matching prediction model is trained by a training dataset that includes multiple vehicle system working reference data and corresponding annotations indicating whether a violation exists; the vehicle system working reference data includes at least one of vehicle system seat belt warning data, vehicle system seat perception data, vehicle system call working data and vehicle system control operation data; The vehicle's operating data is input into the matching prediction model to obtain the output matching probability; When the matching probability is greater than a preset probability threshold, the existing violations and their corresponding matching probabilities are identified as environmental features corresponding to the communication data. When the matching probability is less than the probability threshold, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

[0085] As can be seen, through the above optional embodiments, by determining the matching prediction model corresponding to the violation in the matching model database, inputting the vehicle system working data to obtain the matching probability, and determining the environmental characteristics based on the probability, the system can accurately quantify the violation environmental characteristics based on vehicle system data and model prediction, improve the intelligence and objectivity of environmental feature recognition, and reduce the risk of violation omissions due to single rules.

[0086] As an optional embodiment, the prediction module predicts the specific manner of future driving behavior based on the driving and environmental characteristics corresponding to all communication data, using a recurrent neural network, including: The data sequence is obtained by sorting all communication data from morning to night based on the corresponding upload time points; The corresponding driving features and environmental features are determined as data feature labels for each communication data in the data sequence to obtain the updated labeled data sequence; Based on recurrent neural networks, future driving behavior can be predicted from labeled data sequences.

[0087] As can be seen, through the above optional embodiments, by sorting communication data based on upload time points to obtain a data sequence, and labeling driving features and environmental features as data features to obtain a labeled data sequence, and predicting future driving behavior based on recurrent neural networks, the temporal labeling of historical communication data and the prediction of future behavior are realized, thereby improving the continuity and foresight of the prediction process and reducing the risk of prediction deviation caused by unlabeled time series.

[0088] As an optional embodiment, the prediction module is based on a recurrent neural network and predicts the specific manner of future driving behavior based on a labeled data sequence, including: The labeled data sequence is input into the trained recurrent neural network model to obtain the output future driving behavior; optionally, the recurrent neural network model is trained using a training dataset that includes driving behavior sequences of multiple training vehicles and corresponding T-Box device communication data annotations, driving feature annotations, and environmental feature annotations; the future driving behavior includes multiple driving operations ordered according to their possible time.

[0089] As can be seen, through the above optional embodiments, future driving behavior can be obtained by inputting labeled data sequences into a trained recurrent neural network model, thereby achieving deep prediction based on multi-labeled time series data, improving the accuracy and generalization ability of future driving behavior prediction, and reducing the risk of prediction errors caused by incomplete model input.

[0090] As an optional embodiment, the calculation module calculates the driving risk parameters corresponding to the vehicle to be identified based on future driving behavior and risk accumulation calculation rules in the following specific ways: Determine the current risk parameters corresponding to the vehicle to be identified; optionally, the current risk parameters are calculated based on the driving characteristics corresponding to the communication data using preset risk scoring rules; Based on the risk scoring rules and future driving behavior, calculate the cumulative risk parameters; Calculate the similarity between future driving behavior and the current driving feature set, and calculate the parameter weights proportional to the similarity; optionally, the current driving feature set includes multiple driving features. The modified cumulative risk parameter is obtained by multiplying the parameter weights and the cumulative risk parameter. The sum of the current risk parameters and the corrected cumulative risk parameters is calculated to obtain the driving risk parameters corresponding to the vehicle to be identified.

[0091] As can be seen, through the above optional embodiments, by determining the current risk parameters, calculating the cumulative risk parameters according to the risk scoring rules and future driving behavior, and calculating the weight to correct the cumulative risk parameters based on the similarity between future behavior and current driving characteristics, the driving risk parameters are finally obtained by calculating the sum of the current risk and the corrected cumulative risk. This achieves a fusion and quantitative assessment of current and future risks, improves the comprehensiveness and dynamism of driving risk parameters, and reduces the risk of underestimation due to failure to consider future behavior.

[0092] Example 3 Please see Figure 3 , Figure 3 This is another T-Box data-driven dynamic assessment system for vehicle driving risks disclosed in the embodiments of the present invention. Figure 3 The described T-Box data-driven dynamic assessment system for vehicle driving risks is applied in a data processing system / data processing equipment / data processing server (wherein, the server includes a local processing server or a cloud processing server). For example... Figure 3 As shown, the T-Box data-driven vehicle driving risk dynamic assessment system may include: Memory 301 storing executable program code; Processor 302 coupled to memory 301; The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the T-Box data-driven dynamic assessment method for vehicle driving risks described in Embodiment 1.

[0093] Example 4 This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps of the T-Box data-driven dynamic assessment method for vehicle driving risks described in Embodiment 1.

[0094] Example 5 This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the T-Box data-driven dynamic assessment method for vehicle driving risks described in Embodiment 1.

[0095] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0097] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware.

[0098] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0099] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0102] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0103] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0104] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0105] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0106] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0107] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0108] Finally, it should be noted that the T-Box data-driven dynamic assessment method and system for vehicle driving risks disclosed in the embodiments of this invention are merely preferred embodiments of the invention and are only used to illustrate the technical solutions of the invention, not to limit it. Although the invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this invention.

Claims

1. A T-Box data-driven method for dynamic assessment of vehicle driving risk, characterized in that, The method includes: Acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated; Based on data analysis algorithms, the driving characteristics and environmental characteristics corresponding to each piece of communication data are determined. Based on the driving and environmental characteristics corresponding to all the aforementioned communication data, future driving behavior is predicted using a recurrent neural network. Based on the future driving behavior and the risk accumulation calculation rules, the driving risk parameters corresponding to the vehicle to be identified are calculated.

2. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 1, characterized in that, The method of determining the driving characteristics and environmental characteristics corresponding to each piece of communication data based on data analysis algorithms includes: For each of the aforementioned communication data, based on character recognition rules, the vehicle driving data, environmental sensing data, and vehicle system operating data in the communication data are identified; Based on the vehicle driving data, analyze the driving characteristics corresponding to the communication data; Based on the environmental sensing data and the vehicle system operating data, the environmental characteristics corresponding to the communication data are analyzed.

3. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 2, characterized in that, The step of analyzing the driving characteristics corresponding to the communication data based on the vehicle driving data includes: Based on the vehicle driving data, determine the vehicle acceleration, vehicle direction of travel, and vehicle driving operation of the vehicle to be evaluated corresponding to the communication data; Based on the preset threshold rules for the three urgent behaviors, as well as the vehicle acceleration and the vehicle's direction of travel, it is determined whether the vehicle to be evaluated exhibits the three urgent behaviors, and a first judgment result is obtained. When the first determination result is yes, the vehicle driving operation, the vehicle acceleration, and the corresponding three emergency behavior types are determined as the driving characteristics corresponding to the communication data; the three emergency behavior types are rapid acceleration, sharp turning, or rapid deceleration. If the first determination result is negative, the vehicle driving operation is determined as the driving feature corresponding to the communication data.

4. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 2, characterized in that, The step of analyzing the environmental characteristics corresponding to the communication data based on the environmental sensing data and the vehicle system operating data includes: The environmental sensing data is input into the trained in-vehicle activity prediction model to obtain the output predicted in-vehicle behavior; the in-vehicle activity prediction model is trained using a training dataset that includes multiple training environmental sensing data and corresponding in-vehicle behavior annotations. Determine whether the predicted in-vehicle behavior includes any preset violations to obtain a second determination result; the violations include at least one of the following: not wearing a seat belt, having children but not using a child seat, the driver making a phone call, overloading the vehicle, and improperly placing the seats in the vehicle. When the second judgment result is yes, based on the matching model corresponding to the existing violation, the environmental characteristics corresponding to the communication data are determined according to the vehicle system working data; If the second determination result is negative, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

5. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 4, characterized in that, The matching model based on existing violations determines the environmental characteristics corresponding to the communication data based on the vehicle system's operating data, including: The matching prediction model corresponding to the existing violation is determined in the preset matching model database; the matching prediction model is trained by a training dataset including multiple vehicle system working reference data and corresponding annotations indicating whether the violation exists; the vehicle system working reference data includes at least one of vehicle system seat belt warning data, vehicle system seat perception data, vehicle system call working data and vehicle system control operation data. The vehicle system's operating data is input into the matching prediction model to obtain the output matching probability; When the matching probability is greater than a preset probability threshold, the existing violation and the corresponding matching probability are determined as the environmental characteristics corresponding to the communication data. When the matching probability is less than the probability threshold, the environmental sensing data is determined as the environmental feature corresponding to the communication data.

6. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 1, characterized in that, The step of predicting future driving behavior based on driving and environmental features corresponding to all the communication data, using a recurrent neural network, includes: The data sequence is obtained by sorting all the communication data from morning to night based on the corresponding upload time points; The corresponding driving features and environmental features are determined as data feature labels for each communication data in the data sequence to obtain an updated labeled data sequence; Based on a recurrent neural network, future driving behavior is predicted according to the labeled data sequence.

7. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 6, characterized in that, The method of predicting future driving behavior based on the labeled data sequence using a recurrent neural network includes: The labeled data sequence is input into a trained recurrent neural network model to obtain the output future driving behavior; the recurrent neural network model is trained using a training dataset that includes driving behavior sequences of multiple training vehicles and corresponding T-Box device communication data annotations, driving feature annotations, and environmental feature annotations; the future driving behavior includes multiple driving operations ordered according to their possible time.

8. The T-Box data-driven dynamic assessment method for vehicle driving risk according to claim 1, characterized in that, The step of calculating the driving risk parameters corresponding to the vehicle to be identified based on the future driving behavior and the risk accumulation calculation rules includes: Determine the current risk parameters corresponding to the vehicle to be identified; the current risk parameters are calculated based on the driving characteristics corresponding to the communication data using a preset risk scoring rule; Based on the risk scoring rules and the future driving behavior, calculate the cumulative risk parameters; Calculate the similarity between the future driving behavior and the current driving feature set, and calculate the parameter weights that are proportional to the similarity; the current driving feature set includes multiple driving features. The modified cumulative risk parameter is obtained by multiplying the parameter weights and the cumulative risk parameter. The sum of the current risk parameter and the corrected cumulative risk parameter is calculated to obtain the driving risk parameter corresponding to the vehicle to be identified.

9. A T-Box data-driven dynamic assessment system for vehicle driving risks, characterized in that, The system includes: The acquisition module is used to acquire multiple communication data uploaded by the T-Box device of the vehicle to be evaluated; The determination module is used to determine the driving characteristics and environmental characteristics corresponding to each piece of communication data based on data analysis algorithms; The prediction module is used to predict future driving behavior based on a recurrent neural network, according to the driving characteristics and environmental characteristics corresponding to all the communication data. The calculation module is used to calculate the driving risk parameters corresponding to the vehicle to be identified based on the future driving behavior and the risk accumulation calculation rules.

10. A T-Box data-driven dynamic assessment system for vehicle driving risk, characterized in that, The system includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the T-Box data-driven dynamic assessment method for vehicle driving risks as described in any one of claims 1-8.

Images