Accident early warning analysis method and system based on driving data

Abstract

The invention discloses an accident early warning analysis method and system based on driving data, and relates to the technical field of accident early warning analysis, by extracting a road condition complexity coefficient Cen, a sight distance chaos coefficient Cvs and a dynamic interference coefficient Cdn, the system not only considers the current state, but also analyzes the dynamic interference characteristics of the current state when modeling the external environment, so that the accident early warning analysis accuracy is improved. And the system still has high reliability in the scenes of rain and fog, night, congestion and the like. The comprehensive score value Srisk is not single binary judgment any more, multi-factor fusion interpretation is carried out in combination with information such as operation disturbance, posture deviation and fatigue indexes, a response result WIN with grading characteristics is output, and multi-level early warning from voice reminding to active intervention can be achieved. And the dynamic feedback optimization module constructs a risk response parameter psi stab according to a score fluctuation trend, dynamically adjusts a weight coefficient omega j of the multi-modal model, generates a new coefficient n omega, and realizes self-parameter adjustment and adaptive reinforcement learning.

Description

Technical Field

[0001] The present invention relates to the technical field of accident warning analysis, and in particular to an accident warning analysis method based on driving data and a system thereof. Background Art

[0002] In active safety control, modules such as vehicle collision warning, lane keeping assistance, and pedestrian recognition reminder are particularly critical subsystems in advanced driver assistance systems. The accident warning analysis system based on driving data is one of the key core modules embedded in this type of ADAS architecture. It identifies potential risk factors through real-time analysis of multi-modal information such as vehicle dynamics, environmental perception, and driver behavior. It is the core bridge that promotes autonomous driving from perception to prediction and then to the closed control loop.

[0003] Most current warning systems rely on single-scenario modeling or hard-rule triggering mechanisms, making it difficult to accurately model the coupling relationship between complex driving behaviors and dynamic environments. For example, in the FCW system, fixed time threshold judgments are generally used, but the driver's operating intentions or environmental complexity at the time are not taken into account, which can easily lead to false alarms or missed alarms; PCW's recognition accuracy decreases when facing multi-posture pedestrians, especially in night scenes, where the recognition accuracy of traditional models is heavily dependent on lighting conditions; the HMW system also has blind spots that cannot monitor distant targets. The "modal split", "behavior isolation" and "prediction lag" phenomena in these existing systems make the triggering of warning signals lack flexibility and accuracy, making it difficult to adapt to the increasingly complex and changing road traffic environment.

[0004] The fundamental reason behind these current situations is that, on the one hand, the warning system lacks a unified modeling mechanism for driving behavior, environmental changes, and risk patterns, which leads to a narrow understanding of the system, especially the lack of the ability to predict the evolution of behavior trends; on the other hand, the use of driving data in traditional systems only stays at the surface threshold judgment, without the introduction of deep feature extraction and time series prediction and other intelligent analysis methods, and cannot dynamically adapt to the driver's state changes and environmental interference. These defects may lead to serious consequences in actual operation. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an accident warning analysis method based on driving data and a system thereof, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an accident warning analysis system based on driving data, including a data acquisition and preprocessing module, an abnormal behavior recognition module, an environmental complexity recognition module, a multi-mode risk fusion assessment module, a comprehensive assessment and warning decision module and a dynamic feedback optimization module; The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; The comprehensive assessment and early warning decision module comprehensively interprets the acquired risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

[0007] Preferably, the data acquisition and preprocessing module includes a multi-source data acquisition unit and a standardized preprocessing unit; The multi-source data acquisition unit collects the multi-dimensional driving data of the car, including the vehicle vertical vibration frequency Fv, the driver's steering action micro-response time Ft, the visual range image ambiguity Fb, the front target micro-speed change rate Fp and the driver's blink rate Fe, and performs fitting to obtain the original data set WQ; Among them, the vehicle vertical vibration frequency Fv and the vehicle lateral acceleration instantaneous value alt are collected by the IMU sensor built into the driving recorder; The driver's steering action micro-response time Ft is collected through the steering angle sensor; The blur of the sight image Fb and the distance to the front vehicle df are collected by the high-definition camera in the driving recorder; The micro-speed change rate Fp of the target ahead is collected by millimeter-wave radar; The driver’s blink rate Fe is collected through the in-car camera; The standardization preprocessing unit removes outliers and normalizes the original data set WQ to obtain the driving data set WS; Outlier removal includes processing the original data set WQ using the triple standard deviation method to remove outliers; Normalization processing uses the maximum and minimum normalization method to process the original data set WQ to obtain the driving data set WS; The driving dataset WS is obtained by the following formula: ; Wherein, WSb represents the b-th data item in the driving dataset WS, WQb represents the b-th data item in the original dataset WQ, minWQb represents the valley value of the b-th data item in the original dataset WQ, and maxWQb represents the peak value of the b-th data item in the original dataset WQ.

[0008] Preferably, the abnormal behavior identification module includes an abnormal behavior feature extraction unit and an abnormal feature fusion unit; The abnormal behavior feature extraction unit extracts abnormal behavior features from the driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos; The driving behavior fluctuation index Dar is obtained by the following formula: ; Where Tp is the standardized reference value of the speed change rate of the target ahead, and Te is the standardized reference value of the driver's blink rate. The operating rhythm imbalance index Dos is obtained by the following formula: ; In the formula, alt represents the instantaneous value of the vehicle's lateral acceleration, palt represents the average value of the lateral acceleration, and Fv represents the vertical vibration frequency of the vehicle; The abnormal feature fusion unit fuses the driving behavior fluctuation index Dar and the operation rhythm imbalance index Dos to calculate and obtain the abnormal behavior parameter YC; The abnormal behavior parameter YC is obtained by the following formula: ; In the formula, They respectively represent the preset weight values ​​of the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos.

[0009] Preferably, the environment complexity identification module includes a road condition complexity analysis unit and a complexity parameter fusion unit; The road condition complexity analysis unit extracts road condition complexity features from the driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn; The road complexity coefficient Cen is obtained by the following formula: ; In the formula, Fb represents the blur of the sight image, Fp represents the rate of change of the micro-speed of the target in front, df represents the distance of the front vehicle, which is obtained from the driving recorder, and alt represents the instantaneous value of the vehicle's lateral acceleration; The line-of-sight confusion coefficient Cvs is obtained by the following formula: ; In the formula, E represents a constant to prevent division by zero error, and Fb represents the blurriness of the viewing image; The dynamic interference coefficient Cdn is obtained by the following formula: ; Where, Ft represents the micro-response time of the driver's steering action, and Fe represents the driver's blinking rate; The complexity parameter fusion unit fuses the acquired road condition complexity coefficient Cen, sight range confusion coefficient Cvs and power interference coefficient Cdn to calculate and obtain the complexity parameter YZ; The complexity parameter YZ is obtained by the following formula: ; In the formula, They respectively represent the preset weight values ​​of the road condition complexity coefficient Cen, the sight distance confusion coefficient Cvs and the dynamic interference coefficient Cdn.

[0010] Preferably, the multimodal risk fusion assessment module includes a risk score fusion unit and a multimodal risk aggregation unit; The risk score fusion unit fuses the acquired abnormal behavior parameter YC and complexity parameter YZ to obtain the warning risk set R; The early warning risk set R is obtained by the following formula: ; In the formula, R YC represents the score of abnormal behavior parameters, R YZ represents the score of the complexity parameter, maxYC represents the maximum value of the abnormal behavior parameter, and maxYZ represents the maximum value of the complexity parameter.

[0011] Preferably, the multimodal risk aggregation unit calculates the warning risk set R based on the obtained warning risk set R by using the multimodal risk aggregation model to obtain the risk score Rag, and compares it with the preset risk threshold TRa to determine the risk status; The risk score Rag is obtained by the following formula: ; Where Z represents the normalization factor, N represents the number of modes in the early warning risk set R, including the score R of the abnormal behavior parameter YC and the score R of the complexity parameter YZ, Rj represents the jth mode in the warning risk set R, ln represents the logarithmic function, represents the weight coefficient of the jth mode, and Pj represents the occurrence probability of the jth mode; The risk status is obtained by matching: When the risk score Rag ≤ the risk threshold TRa, it means there is no risk; When the risk score Rag>risk threshold TRa, it indicates that there is a risk.

[0012] Preferably, the comprehensive assessment and early warning decision module includes a risk scoring intelligent interpretation unit and a risk determination and response control unit; The risk score intelligent interpretation unit performs a comprehensive interpretation based on the obtained risk score Rag, combined with the blur of the sight image Fb, the driver's steering action micro-response time Ft and the driver's blink rate Fe, and calculates the comprehensive score value Srisk; The comprehensive score Srisk is obtained by the following formula: ; In the formula, ln represents the logarithmic function, Fb represents the blur of the viewing distance image, represents the normalized index of driver reaction delay, represents the blink rate deviation rate; Normalized driver reaction delay index Obtained by the following formula: ; Where Ft represents the micro-response time of the driver's steering action, and Tt represents the reference steering response time. Blink rate offset Obtained by the following formula: ; Where pFe is the normal reference rate of the driver's blinking.

[0013] Preferably, the risk determination and response control unit compares the obtained comprehensive score value Srisk with a preset comprehensive threshold TSr, and triggers different levels of early warning measures according to the determination result; The judgment result is obtained by matching in the following ways: When the comprehensive score value Srisk ≤ the comprehensive threshold TSr, no alarm is triggered; When the comprehensive score value Srisk>comprehensive threshold TSr, an alarm is triggered, the alarm is divided into three levels, and the corresponding measures WIN are generated; The corresponding measure WIN is obtained by the following formula: ; Wherein, In1 and In2 represent the adjustable increment range of the threshold, and In2 is greater than In1.

[0014] Preferably, the dynamic feedback optimization module performs a fluctuation trend analysis on the comprehensive score value Srisk within a fixed period, establishes a quantitative stability evaluation index: the risk response parameter Ψstab, and presets a false alarm threshold Tst to judge the false alarm state; The risk response parameter Ψstab is obtained by the following formula: ; Where m represents the number of ratings in the observation period, Srisk,k represents the comprehensive rating value of the kth period, Srisk,k-1 represents the comprehensive rating value of the k-1th period, Δtk represents the time interval of the kth rating change, and ηk represents the kth rating error rate; The false alarm status is obtained by matching the following methods: When the risk response parameter Ψstab ≤ the false alarm threshold Tst, it means that the model status is normal, the false alarm rate is normal, and the current configuration is maintained; When the risk response parameter Ψstab>false alarm threshold Tst, it means that the model state is abnormal, the score fluctuates, the false alarm rate is abnormal, and the coefficient ωj of the multimodal risk aggregation model is readjusted to obtain a new coefficient nω; The new coefficient nω is obtained by the following formula: ; Wherein, μ represents the adjustment coefficient and E represents the constant.

[0015] An accident warning analysis method based on driving data comprises the following steps: Step 1: The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; Step 2: The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; Step 3: The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; Step 4: The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; Step 5: The comprehensive assessment and early warning decision module comprehensively interprets the obtained risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; Step 6: The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

[0016] The present invention provides an accident warning analysis method and system based on driving data, which has the following beneficial effects: (1) When the system is running, by extracting the road complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, the system of the present invention not only considers the current state of the modeling of the external environment, but also analyzes its dynamic interference characteristics, so that the system still has high reliability in rain, fog, night, congestion and other scenes. The comprehensive score value Srisk is no longer a single binary judgment, but a multi-factor fusion interpretation combined with information such as operation disturbance, posture deviation, fatigue index, etc., and outputs a response result WIN with hierarchical characteristics, which can realize multi-level warnings from voice reminders to active intervention.

[0017] The dynamic feedback optimization module constructs the risk response parameter Ψstab according to the score fluctuation trend, and dynamically adjusts the multimodal model weight coefficient ωj to generate a new coefficient nω, realizing the model's self-tuning and adaptive reinforcement learning, effectively solving the problem of the system's "weak adaptability" in different scenarios. The triple fusion strategy of driving behavior + environmental perception + aggregation algorithm is used to achieve identification closer to the real driving risk scenario, improve the system's warning accuracy and sensitivity under complex conditions, and effectively reduce false alarms and missed alarms.

[0018] (2) Adopting an explicit linear fusion structure, each feature is given a reasonable weight, so as to achieve modeling of the differential contribution of indicators of different dimensions. Compared with "average addition" or "hard threshold comparison", this method can better reflect the sensitivity and importance of each indicator in different situations, and enhance the decision-making flexibility and interpretability of the system.

[0019] Combining volatility analysis with rhythm imbalance analysis, the system can not only identify obvious operational anomalies such as sudden braking and yaw, but also hidden dangerous behaviors such as frequent blinking, delayed response, and slight corrections, achieving earlier and more hidden risk prediction. By integrating multiple factors such as line of sight quality, target dynamics, acceleration disturbance, and driver status, the system can build a comprehensive judgment model for complex environments, especially in extreme driving situations such as rain and fog, high speed, and night, and still maintain stable complexity recognition performance.

[0020] (3) By setting up two-level functional modules, namely the risk score fusion unit and the multimodal risk aggregation unit, a closed-loop process is realized, from the normalization fusion of the abnormal behavior parameter YC and the environmental complexity parameter YZ to the weighted calculation aggregation and then to the risk status output. This breaks through the problem that the traditional system relies only on a single threshold trigger and a rough judgment mechanism, and builds a fully structured risk fusion and response mechanism.

[0021] Through standardized scoring and probability weighting mechanisms, parameters with obvious physical dimension differences can be uniformly modeled, which improves the system's understanding accuracy of multimodal risk signals and significantly reduces fusion errors and judgment offsets. By introducing the modal occurrence probability Pj, the system can adjust the influence of different modal risks in real time according to the actual risk evolution, realize dynamic capture of the "dominant risk factors" in complex scenarios, and improve the agility of overall risk perception.

[0022] (4) Through the intelligent interpretation unit of risk scoring, the blur of the visual range image, the driver's reaction delay index and the blink offset index are introduced on the basis of the original risk score to construct a comprehensive score value of multi-factor fusion. This structure enables the system to rise from a single numerical judgment to a comprehensive interpretation of multimodal behavior and status, solving the problem that the scoring results of traditional models lack semantic support and weak explanatory power. In the scoring calculation process, the driver's micro-response time and physiological fatigue state, as well as the current visual environment quality, are combined to realize the construction of a coordinated scoring method for the three domains of behavior-physiology-environment, making the risk results more contextually understandable and closer to the actual driving status. This approach effectively makes up for the modeling shortcomings of the existing warning system that lacks subjective factors of the driver. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a schematic diagram of a flow chart of an accident warning analysis system based on driving data according to the present invention; Figure 2 A schematic diagram of the steps of an accident warning analysis method based on driving data of the present invention; Figure 3 A schematic diagram of the system flow of the present invention for determining whether a model is normal; Figure 4 A line graph of abnormal behavior parameters of the present invention; Figure 5 This is a schematic structural diagram of a driving recorder used in the present invention. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] Example 1 The present invention provides an accident warning analysis system based on driving data, please refer to Figures 1 to 5 , including data acquisition and preprocessing module, abnormal behavior recognition module, environmental complexity recognition module, multi-mode risk fusion assessment module, comprehensive assessment and early warning decision module and dynamic feedback optimization module; The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; The comprehensive assessment and early warning decision module comprehensively interprets the acquired risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

[0026] In this embodiment, the two-dimensional fusion of abnormal behavior parameter YC and complexity parameter YZ is introduced to construct the early warning risk set RRR, and the multimodal risk aggregation model is used for integrated scoring, which effectively avoids the problems of modal fragmentation and local judgment and improves the comprehensive understanding ability of the system. In the abnormal behavior recognition module, by analyzing the driving behavior fluctuation index Dar and the operation rhythm imbalance index Dos, the driver's behavior is trend analyzed instead of instantaneous judgment, so that the system can make time-series predictions for potential risk behaviors and improve the defects of the traditional system prediction lag.

[0027] By extracting the road complexity coefficient Cen, the sight confusion coefficient Cvs and the dynamic interference coefficient Cdn, the system of the present invention not only considers the current state of the modeling of the external environment, but also analyzes its dynamic interference characteristics, so that the system still has high reliability in rain, fog, night, congestion and other scenes. The comprehensive score value Srisk is no longer a single binary judgment, but a multi-factor fusion interpretation combined with information such as operation disturbance, posture deviation, fatigue index, etc., and outputs a response result WIN with hierarchical characteristics, which can realize multi-level warnings from voice reminders to active intervention.

[0028] The dynamic feedback optimization module constructs the risk response parameter Ψstab according to the score fluctuation trend, and dynamically adjusts the multimodal model weight coefficient ωj to generate a new coefficient nω, realizing the model's self-tuning and adaptive reinforcement learning, effectively solving the problem of the system's "weak adaptability" in different scenarios. The triple fusion strategy of driving behavior + environmental perception + aggregation algorithm is used to achieve identification closer to the real driving risk scenario, improve the system's warning accuracy and sensitivity under complex conditions, and effectively reduce false alarms and missed alarms.

[0029] The analysis of abnormal behavior modeling and score change trends enables the system to identify the risk evolution trend in advance, and the early warning can intervene earlier, truly possessing the timeliness characteristics of "prediction + early warning". The environmental complexity recognition module enables the system to cope with complex, changeable and uncertain environments, and is no longer limited to ideal or single scenarios. The risk score fluctuation analysis and dynamic weight optimization mechanism can ensure that the system adjusts the decision logic in time when the score fluctuates abnormally, improving the robustness of the overall risk perception and the stability of the system operation. Through the dynamic feedback mechanism, the system has the ability to automatically optimize the parameter structure according to historical operation data, and can make adaptive adjustments according to different driver habits and road types to achieve customized and personalized safety assistance. From data collection, behavior recognition, environmental modeling, to risk assessment, early warning response, and then to model optimization feedback, a highly coupled, highly closed-loop, and highly controllable intelligent early warning analysis system has been constructed to promote the intelligent assisted driving system to move towards high-level automatic decision-making.

[0030] Example 2 This embodiment is explained in Example 1, please refer to Figure 1 and Figure 5 ,Specifically: the data acquisition and preprocessing module includes a multi-source data acquisition unit and a ,standardized preprocessing unit; The multi-source data acquisition unit collects the multi-dimensional driving data of the car, including the vehicle vertical vibration frequency Fv, the driver's steering action micro-response time Ft, the visual range image ambiguity Fb, the front target micro-speed change rate Fp and the driver's blink rate Fe, and performs fitting to obtain the original data set WQ; Among them, the vehicle vertical vibration frequency Fv and the vehicle lateral acceleration instantaneous value alt are collected by the IMU sensor built into the driving recorder; The driver's steering action micro-response time Ft is collected through the steering angle sensor; The blur of the sight image Fb and the distance to the preceding vehicle df are collected by a high-definition camera in the driving recorder; The micro-speed change rate Fp of the target ahead is collected by millimeter-wave radar; The driver’s blink rate Fe is collected through the in-car camera; The standardization preprocessing unit removes outliers and normalizes the original data set WQ to obtain the driving data set WS; Outlier removal includes processing the original data set WQ using the triple standard deviation method to remove outliers; Normalization processing uses the maximum and minimum normalization method to process the original data set WQ to obtain the driving data set WS; The driving dataset WS is obtained by the following formula: ; Wherein, WSb represents the b-th data item in the driving dataset WS, WQb represents the b-th data item in the original dataset WQ, minWQb represents the valley value of the b-th data item in the original dataset WQ, and maxWQb represents the peak value of the b-th data item in the original dataset WQ.

[0031] In this embodiment, by setting up a multi-source data acquisition unit, five major parameters, namely, the vehicle vertical vibration frequency Fv, the driver's steering action micro-response time Ft, the visual range image fuzziness Fb, the front target micro-speed change rate Fp, and the driver's blink rate Fe, are simultaneously introduced at the data level for the first time. Compared with the traditional single structure that only focuses on vehicle speed or acceleration, the richness and risk description of the original data are greatly improved, and the three-domain acquisition of "behavior-physiology-environment" is realized, providing more expressive data support for subsequent risk modeling. The data acquisition equipment clearly corresponds to a highly adaptable physical sensor, realizes the structural closure and clear source of the acquisition link, avoids data deviation, improves the stability and controllability of the original data, and provides sensor-level support for building a high-reliability model in real scenarios.

[0032] In the standardization preprocessing unit, this embodiment uses the triple standard deviation method to remove abnormal points from the original data set WQWQWQ. This statistical noise filtering mechanism can effectively shield interference factors such as equipment false alarms and sampling jumps. It is more robust than the traditional "average + direct modeling" method and lays the foundation for high-precision modeling. The maximum-minimum normalization method is used to standardize various types of driving data, so that parameters of different physical units can be compared and integrated on a unified scale, which improves the efficiency of model training and the interpretability of risk analysis results, and enhances the system's cross-scenario adaptation capabilities.

[0033] By introducing a multi-source, multi-modal acquisition mechanism, a cross-level, multi-dimensional panoramic data of driving status is constructed, providing high-density, high-dimensional original information support for abnormal behavior analysis and environmental complexity modeling. In the acquisition stage, an outlier removal and normalization mechanism is introduced to effectively eliminate extreme errors, environmental noise and hardware anomalies, significantly improving the stability, effectiveness and fitting accuracy of modeling data. The normalized data set WS allows the model to perform unified risk score calculations based on data of different types and dimensions, avoiding risk judgment errors caused by inconsistent units or numerical scale deviations.

[0034] Example 3 This embodiment is explained in Example 2. Please refer to Figure 1 and Figure 4 ,Specifically: the abnormal behavior recognition module includes an abnormal behavior feature extraction unit and an abnormal feature fusion unit; The abnormal behavior feature extraction unit extracts abnormal behavior features from the driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos; The driving behavior fluctuation index Dar is obtained by the following formula: ; Where Tp is the standardized reference value of the speed change rate of the target ahead, and Te is the standardized reference value of the driver's blink rate. The operating rhythm imbalance index Dos is obtained by the following formula: ; In the formula, alt represents the instantaneous value of the vehicle's lateral acceleration, palt represents the average value of the lateral acceleration, and Fv represents the vertical vibration frequency of the vehicle; The abnormal feature fusion unit fuses the driving behavior fluctuation index Dar and the operation rhythm imbalance index Dos to calculate the abnormal behavior parameter YC, as shown in Table 1; The abnormal behavior parameter YC is obtained by the following formula: ; In the formula, They respectively represent the preset weight values ​​of the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos.

[0035] Specific example: set up 0.6 and 0.4 respectively; Get the driving behavior fluctuation index Dar=0.50; Operation rhythm imbalance index Dos=0.20; Calculate the abnormal behavior parameter YC: ; Table 1: Sample table of abnormal behavior parameter calculation; The environment complexity recognition module includes a road condition complexity analysis unit and a complexity parameter fusion unit; The road condition complexity analysis unit extracts road condition complexity features from the driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn; The road complexity coefficient Cen is obtained by the following formula: ; In the formula, Fb represents the blur of the visual range image, Fp represents the rate of change of the front target micro-speed, df represents the distance to the front vehicle, and alt represents the instantaneous value of the vehicle's lateral acceleration; The line-of-sight confusion coefficient Cvs is obtained by the following formula: ; In the formula, E represents a constant, and Fb represents the blur of the viewing image; The dynamic interference coefficient Cdn is obtained by the following formula: ; Where, Ft represents the micro-response time of the driver's steering action, and Fe represents the driver's blinking rate; The complexity parameter fusion unit fuses the acquired road condition complexity coefficient Cen, sight range confusion coefficient Cvs and power interference coefficient Cdn to calculate and obtain the complexity parameter YZ; The complexity parameter YZ is obtained by the following formula: ; In the formula, They respectively represent the preset weight values ​​of the road condition complexity coefficient Cen, the sight distance confusion coefficient Cvs and the dynamic interference coefficient Cdn.

[0036] In this embodiment, the abnormal behavior identification module and the environment complexity identification module are further structured, and two-level logic of feature extraction unit + fusion unit are set respectively, so that the extraction and fusion process of each type of information is more systematic and precise, effectively avoiding problems such as information dispersion, structural duplication, and isolated judgment, and enhancing the module collaboration efficiency.

[0037] The extraction of abnormal behavior features is no longer limited to driving operation data. It also introduces the front target micro-speed change rate Fp and the driver's blink rate Fe. From the two perspectives of dynamic response of the driving environment and the driver's physiological state, the following are constructed: driving behavior fluctuation index Dar and operation rhythm imbalance index Dos. Compared with the traditional model that only focuses on steering, acceleration and other operational behaviors, this design is more three-dimensional and physiologically relevant, allowing the system to identify potential abnormal states such as fatigue driving, slow operation, and visual field judgment deviation in advance.

[0038] The environmental complexity parameters no longer only consider the traditional distance and traffic density, but propose the following: Road condition complexity coefficient Cen: Considering the image blur Fb, the front target micro-speed change rate Fp, the lateral acceleration alt, and the front vehicle distance df, a three-factor fusion model of driving vision-dynamic behavior-time-space distance is formed; Vision confusion coefficient Cvs: For scenes of image defocus, night, rainy and foggy weather, the dynamic interference assessment of visual blur is strengthened; Dynamic interference coefficient Cdn: driver's steering action micro-response time Ft and driver's blinking rate Fe, introduces a nonlinear modeling perspective of the interaction between driver status and vehicle dynamics. These coefficients are fused through different dimensions to establish a complexity recognition system that is more in line with the fluctuations of actual driving scenarios. In the calculation of abnormal behavior parameters YC and environmental complexity parameters YZ, this embodiment adopts an explicit linear fusion structure to assign reasonable weights to each feature, thereby achieving differential contribution modeling of indicators of different dimensions. Compared with "average addition" or "hard threshold comparison", this method can better reflect the sensitivity and importance of each indicator in different situations, and enhance the decision-making flexibility and interpretability of the system.

[0039] Combining volatility analysis with rhythm imbalance analysis, the system can not only identify obvious operational anomalies such as sudden braking and yaw, but also hidden dangerous behaviors such as frequent blinking, delayed response, and slight corrections, achieving earlier and more hidden risk prediction. By integrating multiple factors such as line of sight quality, target dynamics, acceleration disturbance, and driver status, the system can build a comprehensive judgment model for complex environments, especially in extreme driving situations such as rain and fog, high speed, and night, and still maintain stable complexity recognition performance.

[0040] Example 4 This embodiment is explained in Example 3, please refer to Figure 1 and Figure 3,Specifically: the multimodal risk fusion assessment module includes a risk score fusion unit and a multimodal risk aggregation unit; The risk score fusion unit fuses the acquired abnormal behavior parameter YC and complexity parameter YZ to obtain the warning risk set R; The early warning risk set R is obtained by the following formula: ; In the formula, R YC represents the score of abnormal behavior parameters, R YZ represents the score of the complexity parameter, maxYC represents the maximum value of the abnormal behavior parameter, and maxYZ represents the maximum value of the complexity parameter.

[0041] The multimodal risk aggregation unit calculates the warning risk set R based on the obtained warning risk set R by using the multimodal risk aggregation model to obtain the risk score Rag, and compares it with the preset risk threshold TRa to determine the risk status; The risk score Rag is obtained by the following formula: ; Where Z represents the normalization factor, N represents the number of modes in the early warning risk set R, including the score R of the abnormal behavior parameter YC and the score R of the complexity parameter YZ , Rj represents the jth mode in the warning risk set R, ln represents the logarithmic function, represents the weight coefficient of the jth mode, and Pj represents the occurrence probability of the jth mode; The risk status is obtained by matching: When the risk score Rag ≤ the risk threshold TRa, it means there is no risk; When the risk score Rag>risk threshold TRa, it indicates that there is a risk.

[0042] In this embodiment, by setting up two-level functional modules of risk score fusion unit and multimodal risk aggregation unit, a closed-loop process of abnormal behavior parameter YC and environmental complexity parameter YZ from score normalization fusion to weighted calculation aggregation and then to risk status output is realized, breaking through the problem that traditional systems rely only on single threshold triggering and rough judgment mechanism, and building a fully structured risk fusion and response mechanism.

[0043] Through standardized scoring and probability weighting mechanisms, parameters with obvious physical dimension differences can be uniformly modeled, which improves the system's understanding accuracy of multimodal risk signals and significantly reduces fusion errors and judgment offsets. By introducing the modal occurrence probability Pj, the system can adjust the influence of different modal risks in real time according to the actual risk evolution, realize dynamic capture of the "dominant risk factors" in complex scenarios, and improve the agility of overall risk perception.

[0044] The risk scoring mechanism has been transformed from a "single indicator" to a "weighted fusion + threshold matching" mode, giving the system a more sophisticated risk identification logic, supporting richer response strategies, and further integrating with risk level models (such as mild / moderate / high risk). The multimodal risk aggregation formula structure is open, and the parameters are expandable (such as introducing more risk modes). It supports the access of new modal variables such as fatigue driving scores and attention status scores, giving the system good iterative upgrade and cross-scenario application capabilities. The risk score Rag, as the final fusion judgment result of the system, can be compared with the actual trigger event in the dynamic feedback optimization module. It serves as a basic reference indicator for weight correction and dynamic adjustment of thresholds, which helps to build a closed-loop self-learning early warning system.

[0045] Example 5 This embodiment is explained in Example 4. Please refer to Figure 1 and Figure 3 ,Specifically: The comprehensive assessment and early warning decision module includes the risk scoring intelligent ,interpretation unit and the risk judgment and response control unit; The risk score intelligent interpretation unit performs a comprehensive interpretation based on the obtained risk score Rag, combined with the blur of the sight image Fb, the driver's steering action micro-response time Ft and the driver's blink rate Fe, and calculates the comprehensive score value Srisk; The comprehensive score Srisk is obtained by the following formula: ; In the formula, ln represents the logarithmic function, Fb represents the blur of the viewing distance image, represents the normalized index of driver reaction delay, represents the blink rate deviation rate; Normalized driver reaction delay index Obtained by the following formula: ; Where Ft represents the micro-response time of the driver's steering action, and Tt represents the reference steering response time. Blink rate offset Obtained by the following formula: ; Where pFe is the normal reference rate of the driver's blinking.

[0046] The risk determination and response control unit compares the obtained comprehensive score value Srisk with the preset comprehensive threshold TSr, and triggers different levels of early warning measures according to the judgment results; The judgment result is obtained by matching in the following ways: When the comprehensive score value Srisk ≤ the comprehensive threshold TSr, no alarm is triggered; When the comprehensive score value Srisk>comprehensive threshold TSr, an alarm is triggered, the alarm is divided into three levels, and the corresponding measures WIN are generated; The corresponding measure WIN is obtained by the following formula: ; Wherein, In1 and In2 represent the adjustable increment range of the threshold, and In2 is greater than In1.

[0047] The dynamic feedback optimization module analyzes the fluctuation trend of the comprehensive score value Srisk within a fixed period, establishes a quantitative stability evaluation indicator: the risk response parameter Ψstab, and the preset false alarm threshold Tst to judge the false alarm state; The risk response parameter Ψstab is obtained by the following formula: ; Where m represents the number of ratings in the observation period, Srisk,k represents the comprehensive rating value of the kth period, Srisk,k-1 represents the comprehensive rating value of the k-1th period, Δtk represents the time interval of the kth rating change, and ηk represents the kth rating error rate; The false alarm status is obtained by matching the following methods: When the risk response parameter Ψstab ≤ the false alarm threshold Tst, it means that the model status is normal, the false alarm rate is normal, and the current configuration is maintained; When the risk response parameter Ψstab>false alarm threshold Tst, it means that the model state is abnormal, the score fluctuates, the false alarm rate is abnormal, and the coefficient ωj of the multimodal risk aggregation model is readjusted to obtain a new coefficient nω; The new coefficient nω is obtained by the following formula: ; Wherein, μ represents the adjustment coefficient and E represents the constant.

[0048] In this embodiment, the intelligent interpretation unit for risk scoring is used to introduce the blur of the line of sight image, the driver's reaction delay index and the blink offset index on the basis of the original risk score to construct a comprehensive score value of multi-factor fusion. This structure enables the system to rise from a single numerical judgment to a comprehensive interpretation of multimodal behavior and state, solving the problem that the scoring results of traditional models lack semantic support and weak explanatory power. In the scoring calculation process, the driver's micro-response time and physiological fatigue state, as well as the current visual environment quality, are combined to realize the construction of a collaborative scoring method for the three domains of behavior-physiology-environment, making the risk results more contextually understandable and closer to the actual driving state. This approach effectively makes up for the shortcomings of the existing early warning system in modeling the lack of subjective factors of the driver.

[0049] The risk determination and response control unit of this embodiment is equipped with a three-level early warning trigger mechanism, which can dynamically adjust the response intensity according to the amplitude of the comprehensive score value, and realize a hierarchical control method from "voice prompt" to "visual reminder" to "automatic control". This progressive response mechanism is more human-computer interaction friendly and scene adaptable than the traditional binary trigger model, reduces driving interference, and improves response matching. The system establishes a risk response parameter Ψstab, which is used to measure the changing trend and false alarm status of the score value within a fixed period, and constructs a stability index system for the scoring curve. This mechanism realizes the evaluation and intervention of the performance of the risk scoring system itself for the first time, provides the system with "self-perception" capabilities, and breaks through the technical limitations of traditional models that cannot determine the source of scoring errors.

[0050] The comprehensive score value is calculated by multi-dimensional key parameters, combined with the driver's status and environmental perception, so that the risk score is no longer an isolated value, but has contextual information and behavioral causal relationships, which improves system transparency and driver trust. Through the hierarchical risk judgment mechanism, the system can adapt to the scene requirements under different risk intensities, effectively avoid false alarms and excessive warnings, and improve the system usage experience and the rationality of the response strategy. Through the evaluation of the score fluctuation trend and the judgment of the false alarm rate, the system can timely identify the abnormal performance of the scoring model, thereby avoiding the output of misjudgment results under abnormal scoring conditions, and significantly enhancing the robustness and reliability of the system. When the scoring system has an error offset, the risk aggregation model can be automatically corrected through a dynamic feedback mechanism to ensure that the system continues to maintain high-performance operation under different driving habits and environmental conditions, and has strong generalization and intelligent evolution capabilities.

[0051] Example 6 An accident warning analysis method based on driving data, please refer to Figure 2 , specifically: including the following steps: Step 1: The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; Step 2: The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; Step 3: The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; Step 4: The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; Step 5: The comprehensive assessment and early warning decision module comprehensively interprets the obtained risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; Step 6: The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

[0052] In this embodiment, the method systematically divides the driving accident warning process into six steps: data collection → anomaly identification → environmental analysis → risk fusion → warning decision → dynamic optimization, forming a full-process closed-loop design from perception, evaluation to response and feedback. This solves the problems of process fragmentation, single evaluation dimension and missing feedback link in traditional methods.

[0053] In the data collection stage, this method is based on the CV2328 four-in-one driving recorder, integrating multiple types of data such as vehicle dynamics, driver status and external environment to construct a multi-dimensional original data set. This multi-source information integrated collection mode significantly improves the breadth and granularity of the original data, providing a high-dimensional and highly expressive input basis for subsequent models. By extracting driving behavior fluctuation indicators and operating rhythm imbalance indexes, deep modeling of driver control stability and physiological rhythm changes is achieved, which can not only identify explicit anomalies, but also implicit behavioral anomalies, significantly improving the system's ability to predict potential risks early.

[0054] In the risk assessment stage, a two-layer structure is adopted, which first normalizes the fusion and then aggregates the modeling. The risk score is calculated based on the probability of occurrence and weighted contribution of the risk modalities. This avoids the problems of "opaque scoring sources and unbalanced modal influences" in traditional methods and improves the interpretability and logical reliability of the scoring.

[0055] A hierarchical output mechanism is introduced in the early warning decision-making stage, and different levels of response measures are intelligently matched according to the comprehensive score value, so as to achieve progressive intervention from mild prompts to system control, greatly improve the human-machine interaction experience and driver acceptance, and avoid driving dependence or boredom caused by excessive intervention. The fusion processing and in-depth evaluation of multi-dimensional risk information are realized to improve the overall risk identification ability and response accuracy of the system; a complete closed-loop intelligent early warning method architecture is constructed, from the bottom-level data collection to the upper-level intelligent control, the whole link is coordinated and efficient, and has strong engineering feasibility and commercial landing value.

[0056] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An accident warning analysis system based on driving data, characterized in that: It includes data collection and preprocessing module, abnormal behavior recognition module, environmental complexity recognition module, multi-mode risk fusion assessment module, comprehensive assessment and early warning decision module and dynamic feedback optimization module; The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; The comprehensive assessment and early warning decision module comprehensively interprets the acquired risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

2. The accident warning analysis system based on driving data according to claim 1, characterized in that: The data acquisition and preprocessing module includes a multi-source data acquisition unit and a standardized preprocessing unit; The multi-source data acquisition unit collects multi-dimensional driving data of the car, including the vehicle vertical vibration frequency Fv, the driver's steering action micro-response time Ft, the visual range image ambiguity Fb, the front target micro-speed change rate Fp, the driver's blink rate Fe, the front vehicle distance df and the vehicle's lateral acceleration instantaneous value alt, and performs fitting to obtain the original data set WQ; Among them, the vehicle vertical vibration frequency Fv and the vehicle lateral acceleration instantaneous value alt are collected by the IMU sensor built into the driving recorder; The driver's steering action micro-response time Ft is collected through the steering angle sensor; The blur of the sight image Fb and the distance to the preceding vehicle df are collected by a high-definition camera in the driving recorder; The micro-speed change rate Fp of the target ahead is collected by millimeter-wave radar; The driver’s blink rate Fe is collected through the in-car camera; The standardization preprocessing unit removes outliers and normalizes the original data set WQ to obtain the driving data set WS; Outlier removal includes processing the original data set WQ using the triple standard deviation method to remove outliers; Normalization processing uses the maximum and minimum normalization method to process the original data set WQ to obtain the driving data set WS; The driving dataset WS is obtained by the following formula: ； Wherein, WSb represents the b-th data item in the driving dataset WS, WQb represents the b-th data item in the original dataset WQ, minWQb represents the valley value of the b-th data item in the original dataset WQ, and maxWQb represents the peak value of the b-th data item in the original dataset WQ.

3. The accident warning analysis system based on driving data according to claim 2, characterized in that: The abnormal behavior recognition module includes an abnormal behavior feature extraction unit and an abnormal feature fusion unit; The abnormal behavior feature extraction unit extracts abnormal behavior features from the driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos; The driving behavior fluctuation index Dar is obtained by the following formula: ； Where Tp represents the standardized reference value of the speed change rate of the target ahead, and Te represents the standardized reference value of the driver's blink rate; The operating rhythm imbalance index Dos is obtained by the following formula: ； In the formula, alt represents the instantaneous value of the vehicle's lateral acceleration, palt represents the average value of the lateral acceleration, and Fv represents the vertical vibration frequency of the vehicle; The abnormal feature fusion unit fuses the driving behavior fluctuation index Dar and the operation rhythm imbalance index Dos to calculate and obtain the abnormal behavior parameter YC; The abnormal behavior parameter YC is obtained by the following formula: ； In the formula, They respectively represent the preset weight values ​​of the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos.

4. The accident warning analysis system based on driving data according to claim 2, characterized in that: The environment complexity recognition module includes a road condition complexity analysis unit and a complexity parameter fusion unit; The road condition complexity analysis unit extracts road condition complexity features from the driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn; The road complexity coefficient Cen is obtained by the following formula: ； In the formula, Fb represents the blur of the visual range image, Fp represents the rate of change of the front target micro-speed, df represents the distance to the front vehicle, and alt represents the instantaneous value of the vehicle's lateral acceleration; The line-of-sight confusion coefficient Cvs is obtained by the following formula: ； In the formula, E represents a constant, and Fb represents the blur of the viewing image; The dynamic interference coefficient Cdn is obtained by the following formula: ； Where, Ft represents the micro-response time of the driver's steering action, and Fe represents the driver's blinking rate; The complexity parameter fusion unit fuses the acquired road condition complexity coefficient Cen, sight range confusion coefficient Cvs and power interference coefficient Cdn to calculate and obtain the complexity parameter YZ; The complexity parameter YZ is obtained by the following formula: ； In the formula, They respectively represent the preset weight values ​​of the road condition complexity coefficient Cen, the sight distance confusion coefficient Cvs and the dynamic interference coefficient Cdn.

5. The accident warning analysis system based on driving data according to claim 1, characterized in that: The multimodal risk fusion assessment module includes a risk score fusion unit and a multimodal risk aggregation unit; The risk score fusion unit fuses the acquired abnormal behavior parameter YC and complexity parameter YZ to obtain the warning risk set R; The early warning risk set R is obtained by the following formula: ； In the formula, R YC represents the score of abnormal behavior parameters, R YZ represents the score of the complexity parameter, maxYC represents the maximum value of the abnormal behavior parameter, and maxYZ represents the maximum value of the complexity parameter.

6. The accident warning analysis system based on driving data according to claim 5, characterized in that: The multimodal risk aggregation unit calculates the warning risk set R based on the obtained warning risk set R by using the multimodal risk aggregation model to obtain the risk score Rag, and compares it with the preset risk threshold TRa to determine the risk status; The risk score Rag is obtained by the following formula: ； Where Z represents the normalization factor, N represents the number of modes in the early warning risk set R, including the score R of the abnormal behavior parameter YC and the score R of the complexity parameter YZ , Rj represents the jth mode in the early warning risk set R, ln represents the logarithmic function, represents the weight coefficient of the jth mode, and Pj represents the occurrence probability of the jth mode; The risk status is obtained by matching: When the risk score Rag ≤ the risk threshold TRa, it means there is no risk; When the risk score Rag>risk threshold TRa, it indicates that there is a risk.

7. The accident warning analysis system based on driving data according to claim 6, characterized in that: The comprehensive assessment and early warning decision module includes a risk scoring intelligent interpretation unit and a risk determination and response control unit; The risk score intelligent interpretation unit performs a comprehensive interpretation based on the obtained risk score Rag, combined with the blur of the sight image Fb, the driver's steering action micro-response time Ft and the driver's blink rate Fe, and calculates the comprehensive score value Srisk; The comprehensive score Srisk is obtained by the following formula: ； In the formula, ln represents the logarithmic function, Fb represents the blur of the viewing distance image, represents the normalized index of driver reaction delay, represents the blink rate deviation rate; Normalized driver reaction delay index Obtained by the following formula: ； Where Ft represents the micro-response time of the driver's steering action, and Tt represents the reference steering response time. Blink rate offset Obtained by the following formula: ； Where pFe is the normal reference rate of the driver's blinking.

8. The accident warning analysis system based on driving data according to claim 7, characterized in that: The risk determination and response control unit compares the obtained comprehensive score value Srisk with the preset comprehensive threshold TSr, and triggers different levels of early warning measures according to the judgment results; The judgment result is obtained by matching in the following ways: When the comprehensive score value Srisk ≤ the comprehensive threshold TSr, no alarm is triggered; When the comprehensive score value Srisk>comprehensive threshold TSr, an alarm is triggered, the alarm is divided into three levels, and the corresponding measures WIN are generated; The corresponding measure WIN is obtained by the following formula: ； Wherein, In1 and In2 represent the adjustable increment range of the threshold, and In2 is greater than In1.

9. The accident warning analysis system based on driving data according to claim 8, characterized in that: The dynamic feedback optimization module analyzes the fluctuation trend of the comprehensive score value Srisk within a fixed period, establishes a quantitative stability evaluation indicator: the risk response parameter Ψstab, and the preset false alarm threshold Tst to judge the false alarm state; The risk response parameter Ψstab is obtained by the following formula: ； Where m represents the number of ratings in the observation period, Srisk,k represents the comprehensive rating value of the kth period, Srisk,k-1 represents the comprehensive rating value of the k-1th period, Δtk represents the time interval of the kth rating change, and ηk represents the kth rating error rate; The false alarm status is obtained by matching the following methods: When the risk response parameter Ψstab ≤ the false alarm threshold Tst, it means that the model status is normal, the false alarm rate is normal, and the current configuration is maintained; When the risk response parameter Ψstab>false alarm threshold Tst, it means that the model state is abnormal, the score fluctuates, the false alarm rate is abnormal, and the coefficient ωj of the multimodal risk aggregation model is readjusted to obtain a new coefficient nω; The new coefficient nω is obtained by the following formula: ； Wherein, μ represents the adjustment coefficient and E represents the constant.

10. An accident warning analysis method based on driving data, applied to an accident warning analysis system based on driving data as claimed in any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: The data acquisition and preprocessing module collects the multi-dimensional driving data of the car through the CV2328 four-in-one driving recorder and sensors, fits it into the original data set WQ, and performs preprocessing to obtain the driving data set WS; Step 2: The abnormal behavior recognition module extracts abnormal behavior features from the acquired driving data set WS, including the driving behavior fluctuation index Dar and the operating rhythm imbalance index Dos, and fits them into abnormal behavior parameters YC; Step 3: The environment complexity recognition module extracts the road condition complexity features from the acquired driving data set WS, including the road condition complexity coefficient Cen, the sight range confusion coefficient Cvs and the dynamic interference coefficient Cdn, and fits them into the complexity parameter YZ; Step 4: The multi-modal risk fusion assessment module fuses the acquired abnormal behavior parameter YC and complexity parameter YZ with risk scores to obtain the warning risk set R, and calculates the warning risk set R using the multi-modal risk aggregation model to obtain the risk score Rag; Step 5: The comprehensive assessment and early warning decision module comprehensively interprets the obtained risk score Rag, obtains the comprehensive score value Srisk, issues an early warning and generates corresponding measures WIN; Step 6: The dynamic feedback optimization module interprets the changing trend of the comprehensive score value Srisk within a fixed period, obtains the risk response parameter Ψstab, and adjusts the coefficient ωj of the multimodal risk aggregation model to obtain a new coefficient nω.

Images