Internet of things based vehicle-road cooperation complex weather traffic safety early warning system and method

Abstract

This invention relates to the field of traffic control technology, and discloses an IoT-based vehicle-road cooperative traffic safety early warning system and method for complex weather. By deeply integrating multi-source data and making intelligent decisions based on a vehicle-road cloud cooperative platform, it accurately identifies and rapidly responds to micro-risks such as icing and fog, creating a closed-loop linkage between dynamic speed limits, rear-end collision prevention warnings, and active illuminated road markings. This significantly reduces the risk of accidents such as skidding and chain-reaction rear-end collisions in complex weather conditions. Simultaneously, through integrated cooperative control, it provides drivers with coherent, prioritized, and cognitively consistent guidance information, greatly reducing their information processing load in adverse environments. Furthermore, an intelligent energy efficiency management strategy based on real-time traffic conditions and risk levels ensures that roadside facilities such as illuminated road markings and warning devices operate in optimal power consumption modes, guaranteeing high reliability service around the clock while achieving long-term economic efficiency and sustainability of the system.

Description

Technical Field

[0001] This invention relates to the field of traffic control technology, and in particular to an Internet of Things-based vehicle-road cooperative traffic safety early warning system and method for complex weather. Background Technology

[0002] With the rapid development of intelligent transportation systems, vehicle-road cooperative technology has become a key direction for improving road safety and efficiency. Especially in complex weather conditions such as rain, fog, snow, and ice, low visibility and a sharp drop in road surface adhesion coefficient can easily trigger traffic accidents, creating an urgent need for real-time, accurate traffic safety early warning and guidance systems. Existing technical solutions generally suffer from core defects when dealing with complex weather conditions, including low system integration, limited dynamic response and collaborative control capabilities, insufficient fusion of multi-source heterogeneous data, and inadequate efficiency and reliability of active lighting facilities. This results in untimely, inaccurate, and uncoordinated early warning information, high cognitive load on drivers, and ultimately an inability to effectively curb traffic accidents under complex weather conditions, especially the high incidence of chain-reaction rear-end collisions. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide an Internet of Things-based vehicle-road cooperative complex weather traffic safety early warning system and method, which can deeply integrate perception data, intelligent collaborative decision-making, and realize all-weather proactive traffic safety early warning through multi-terminal integrated linkage.

[0004] The first aspect of this invention discloses an Internet of Things-based vehicle-road cooperative traffic safety early warning system for complex weather conditions. The system includes an instruction execution module; the instruction execution module includes a dynamic speed limit display unit, a driving warning and guidance unit, and an actively luminous traffic marking unit; the system also includes a vehicle-road cooperative control module; wherein the vehicle-road cooperative control module specifically includes: The sensing unit is used to collect multi-source data through multi-source sensors; the multi-source sensors include sensors at the roadside end, and the collected data includes road surface condition data, traffic flow data, meteorological data, and vehicle identification data of specific types; The data fusion unit is used to receive and fuse the multi-source data to generate first fused data; The instruction generation unit is used to generate collaborative control instructions based on the first fused data, and send the collaborative control instructions to the instruction execution module; The instruction execution module is used to receive the collaborative control instruction and execute the collaborative control instruction in coordination with the dynamic speed limit display unit, the driving warning guidance unit, and the active luminous traffic marking unit.

[0005] Furthermore, the instruction generation unit has a built-in collaborative strategy rule base; The process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction.

[0006] Furthermore, the process of determining the dominant risk scenario based on the first fused data includes: The risk type is determined based on the first fused data; Based on the spatial distribution of sensor data, risk distribution types are identified; the risk distribution types include point-like, segment-like, and region-like distributions. The dominant risk scenario is determined based on the risk type and risk distribution type.

[0007] Furthermore, the data collected by the multi-source sensors at the roadside end also includes vehicle trajectory data; the multi-source sensors also include multi-source sensors at the vehicle end for collecting vehicle-end data; the vehicle-end data includes driver vehicle operation data; The vehicle-road cooperative module also includes a driving behavior analysis unit, which is used to calculate a dynamic behavior index reflecting the driver's state based on the vehicle trajectory data and vehicle operation data.

[0008] Furthermore, the instruction generation unit is also used to determine high-risk driving state vehicles based on the dynamic behavior index, and to adjust the intensity and content of the instructions issued to high-risk driving state vehicles.

[0009] Furthermore, the specific type of vehicle includes at least one of school buses, public buses, large passenger vehicles, trucks, and dangerous goods transport vehicles; The vehicle-road cooperative control module also configures different safety parameter models for different specific types of vehicles, and when generating cooperative control commands based on the first fused data, it also includes: Based on specific vehicle identification information and corresponding safety parameter models, differentiated instructions are generated for specific types of vehicles entering high-risk road sections.

[0010] Furthermore, the vehicle-road cooperative control module also includes an energy efficiency management unit; The energy efficiency management unit determines the working power consumption mode of each unit in the instruction execution module at different times based on the first fused data and the operation content to be performed by each unit in the instruction execution module; the working power consumption mode includes sleep mode, duty mode, enhanced mode and full power warning mode.

[0011] Furthermore, the energy efficiency management unit is also used to obtain real-time energy status information of each unit of the instruction execution module; the instruction generation unit is also used to adjust the parameters of the collaborative control instruction according to the energy status information.

[0012] The second aspect of this invention discloses a vehicle-road cooperative method for traffic safety warning in complex weather based on the Internet of Things. This method is applied to the system disclosed in the first aspect, and includes: Multi-source data is collected through multi-source sensors; the multi-source sensors include roadside sensors, and the collected data includes road surface condition data, traffic flow data, meteorological data, and specific type vehicle identification data; The multi-source data are fused to generate the first fused data; Generate collaborative control instructions based on the first fused data; The coordinated control commands are executed in concert by the dynamic speed limit display unit, the driving warning and guidance unit, and the active luminous traffic marking unit.

[0013] Furthermore, the process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention deeply integrates multi-source data such as road conditions, traffic flow, weather, vehicle type, and vehicle behavior, and makes intelligent decisions based on a vehicle-road-cloud collaborative platform. It accurately identifies and rapidly responds to micro-risks such as icing and fog, creating a closed-loop linkage between dynamic speed limits, rear-end collision warnings, and active illuminated road markings. This significantly reduces the risk of accidents such as skidding and chain-reaction collisions in complex weather conditions. Simultaneously, through integrated collaborative control, it provides drivers with coherent, prioritized, and cognitively consistent guidance information, greatly reducing their information processing load in adverse environments. Furthermore, an intelligent energy efficiency management strategy based on real-time traffic conditions and risk levels ensures that roadside facilities such as illuminated road markings and warning devices operate in optimal power consumption modes, guaranteeing high reliability service around the clock while achieving long-term economic efficiency and sustainability. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings: Figure 1This is a schematic diagram of the structure of a vehicle-road cooperative complex weather traffic safety early warning system based on the Internet of Things disclosed in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the vehicle-road cooperative control module disclosed in Embodiment 1 of the present invention. Detailed Implementation

[0016] 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.

[0017] Example 1: The first aspect of this invention discloses a vehicle-road cooperative traffic safety early warning system based on the Internet of Things (IoT). Please refer to... Figure 1 , Figure 1 This is a schematic diagram of a vehicle-road cooperative traffic safety early warning system based on the Internet of Things (IoT) disclosed in an embodiment of the present invention. The system includes a vehicle-road cooperative control module and a command execution module. The command execution module includes a dynamic speed limit display unit, a driving warning and guidance unit, and an active luminous traffic marking unit.

[0018] It should be noted that the core hardware structure and basic functional principles of the dynamic speed limit display unit, driving warning and guidance unit, and active luminous traffic marking unit involved in the embodiments of the present invention are all mature technologies known in the art. Specifically, the dynamic speed limit display unit mainly refers to the dot-matrix LED variable information sign that has been commercially available on a large scale. It adopts a modular design, has high brightness, wide viewing angle, and good environmental adaptability, and changes the displayed content (such as numbers, arrows, or simple graphics) by receiving external commands. The driving warning and guidance unit can be composed of independent warning devices (such as LED warning screens and directional sound and light alarms). These devices can execute multiple preset response modes according to the externally input risk level signal, such as using different colors, flashing frequencies, or sound and light combinations to represent low, medium, and high risk levels. The active luminous traffic marking unit mainly refers to a structure composed of pre-encapsulated LED modules combined with specific road surface materials (such as epoxy resin and wear-resistant polymers). Its basic technology focuses on the sealing and waterproofing of LEDs, pressure-resistant and wear-resistant encapsulation, and low-power driving. Active luminescent traffic markings are mainly used to replace or enhance the visual guidance function of traditional reflective markings in low-visibility environments (such as nighttime, fog, rain, and snow). By continuously or in a controlled manner emitting light, they clearly mark key geometric lines such as lane boundaries, road outlines, and turning directions for drivers, thereby compensating for the deficiencies of insufficient ambient light or the sharp decline in reflectivity of traditional reflective markings due to surface contamination, aging, or water accumulation. They are a passive safety infrastructure that improves basic road visibility and ensures that vehicles travel along the correct trajectory. Their core lies in providing stable and reliable visual references for road geometry information that do not depend on vehicle lights.

[0019] The core of this invention does not lie in fundamentally modifying the hardware structure of the aforementioned individual units, but rather in placing these three originally independently configured execution units, which passively respond to single signals, under a unified intelligent decision-making and collaborative control framework through the proposed vehicle-road cooperative control module. Based on the fusion and real-time analysis of multi-source heterogeneous data, the vehicle-road cooperative control module generates and issues a highly coordinated instruction package in time, space, and logic, dynamically coordinating the working modes, parameters, and priorities of the three units, thereby generating a synergistic and effective safety early warning effect at the system level.

[0020] Please see Figure 2 , Figure 2 This is a schematic diagram of the vehicle-road cooperative control module disclosed in an embodiment of the present invention. The vehicle-road cooperative control module specifically includes: The sensing unit is used to collect multi-source data through multi-source sensors; the multi-source sensors include sensors at the roadside end, and the collected data includes, but is not limited to, road surface condition data, traffic flow data, meteorological data, and vehicle identification data of specific types.

[0021] The data fusion unit is used to receive and fuse the multi-source data to generate the first fused data.

[0022] The instruction generation unit is used to generate collaborative control instructions based on the first fused data and send the collaborative control instructions to the instruction execution module.

[0023] The instruction execution module is used to receive the collaborative control instruction and execute the collaborative control instruction in coordination with the dynamic speed limit display unit, the driving warning guidance unit, and the active luminous traffic marking unit.

[0024] Specifically, in this embodiment of the invention, the data fusion unit first performs spatiotemporal normalization preprocessing on the raw data streams uploaded by multiple source sensors, uses high-precision timing signals to assign a unified timestamp to all data, and maps all detection targets (such as vehicle positions and slippery road surfaces) to the same high-precision digital map coordinate system based on pre-calibrated sensor geographic coordinates and orientation parameters. At the same time, it converts various types of data (such as pixel coordinates of images, polar coordinates of radar, and scalar values ​​of meteorological sensors) into internally defined standardized data formats to form a basic data set that is comparable and correlated.

[0025] Subsequently, corresponding feature extraction algorithms are run in parallel for different types of standardized data. These include, but are not limited to, extracting vehicle trajectories and lane occupancy status from preprocessed image sequences using target detection and tracking algorithms; extracting precise velocity vectors, accelerations, and headway of moving targets from radar point cloud data; extracting visibility values, precipitation types, and intensity levels from meteorological data streams; and calculating estimated friction coefficients and icing / water accumulation probabilities from dedicated road surface condition detector signals using existing inversion models based on physical principles or empirical data. The extracted features are then evaluated for confidence and input into a feature association engine. This engine, based on the principle of spatiotemporal consistency, associates and verifies different features describing the same object or spatial location. For example, it matches and complementarily fuses visually recognized vehicle boundaries with radar-detected vehicle positions to generate a fused, credible, full-attribute description of the vehicle target.

[0026] Finally, a comprehensive analysis and spatial structure resolution are performed on the entire feature set with precise spatial labels after association. This process not only evaluates whether a single risk factor holds true, such as visibility below a threshold at a certain point, but more importantly, it analyzes the coexistence and coupling relationships of multiple risk factors in space. For example, it identifies spatially overlapping segments of the three features of low visibility, low friction coefficient, and high traffic density. Through algorithms such as spatial clustering, the spatial distribution patterns of specific feature values ​​are analyzed, and raw geometric and statistical quantities describing their distribution characteristics are output, such as the density of distribution points, the boundaries of spatial range, and the continuous length along the road direction. The final output of the first fused data is a structured data object, which encapsulates the quantified values, spatial locations, confidence levels, correlations between features, and the raw analysis results describing the spatial distribution characteristics of the various risk features that have been fused and confirmed.

[0027] Furthermore, the instruction generation unit has a built-in collaborative strategy rule base; The process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction.

[0028] In this embodiment of the invention, the collaborative strategy rule base built into the aforementioned instruction generation unit is essentially a digital decision-making model that maps multidimensional dominant risk scenarios to specific integrated control strategies. The construction and determination of this rule base integrates, but is not limited to, traffic safety engineering principles, historical accident data statistical analysis, traffic flow theoretical models, and human factors engineering cognitive laws. Specifically, the rule base establishment process first deconstructs and summarizes typical risk-causing scenarios under complex weather conditions to form standardized scenario templates, such as a sudden drop in regional visibility accompanied by dry road surfaces. Each scenario template is associated with a set of control objectives that have been quantitatively deduced and verified through simulation, such as reducing the overall vehicle speed variance or maximizing the safe distance between vehicles. Based on these control objectives, the rule base predefines a corresponding collaborative control mode for each scenario template. This mode defines the execution logic relationship between the three, including the temporal sequence, spatial trigger range association, and intensity coupling ratio, thereby constituting a complete instruction parameter combination package.

[0029] In actual operation, the instruction generation unit, based on the received first fused data, determines the dominant risk scenario in real time through feature matching and logical reasoning. Then, using this scenario as an index, it queries and activates the corresponding preset collaborative control mode in the collaborative strategy rule base. The definition of the priority and parameter combination relationship between instructions in this collaborative control mode is achieved through built-in, computable logical rules. Priority relationships primarily address resource competition and attention guidance issues when different safety information is output. For example, in the initial stage of a sharp decrease in visibility, the mode may stipulate that the active luminous road marking system should be triggered first to quickly outline the road contour with high brightness or a specific color (such as yellow), providing the driver with the most basic spatial positioning reference; this instruction has the highest priority. Simultaneously, the rear-end collision warning device enters a high-sensitivity monitoring state but remains silent or provides a low-level warning to avoid information overload. After the basic road contour is established, the display of speed limit instructions and higher-level rear-end collision risk warnings are gradually strengthened. This sequential rather than concurrent instruction issuance logic is the core manifestation of priority relationships.

[0030] The parameter combination relationship reflects the intelligent linkage and mutual correction between the various units of the instruction execution module at a deeper level. Its determination and manifestation mainly rely on cross-unit parameter association functions or lookup tables. For example, in the cooperative control mode, it may be stipulated that the recommended speed value calculated by the dynamic speed limit instruction will be directly used as a key input parameter for calculating the safe distance threshold in the rear-end collision warning device, ensuring that speed control and distance warning are consistent in safety logic. At the same time, the luminous intensity and flashing frequency parameters of the active luminous markings are not only controlled by the ambient light intensity, but also linked to the real-time traffic density and risk level detected by the rear-end collision warning device. When it is determined that the rear-end collision risk of a certain lane has increased, in addition to activating the dedicated warning device on the roadside of that vehicle, the cooperative control mode will also instruct the active luminous marking unit corresponding to that lane to switch to a higher brightness red breathing flashing mode, thereby binding the abstract high-risk warning with the specific lane-level spatial location, realizing the three-dimensional and precise projection of warning information from the roadside to the road surface. Through this deeply integrated parameter combination rule, the units of the instruction execution module no longer respond in isolation, but form an organic whole that mutually reinforces and complements each other under a unified security goal, ultimately achieving a significant leap in system-level security performance.

[0031] Furthermore, the process of identifying the dominant risk scenario based on the first fused data includes: The risk type is determined based on the first fused data; Based on the spatial distribution of sensor data, risk distribution types are identified; the risk distribution types include point-like, segment-like, and region-like distributions. The dominant risk scenario is determined based on the risk type and risk distribution type.

[0032] In this embodiment of the invention, the dominant risk scenario refers to an integrated situational assessment conclusion with clear spatial attributes and threat levels, the purpose of which is to accurately match and trigger subsequent preset collaborative control modes. Specifically, before determining the dominant risk scenario, it is necessary to determine the risk type, that is, to semantically classify and logically judge the features of the first fused data. Based on the built-in risk classification rules, the key feature values ​​in the first fused data are compared and analyzed. For example, when the visibility value is consistently lower than a preset threshold, it is determined that there is a low visibility risk; when the estimated road friction coefficient is lower than a safety threshold, it is determined that there is a slippery road surface risk; when the headway statistics are generally lower than the critical value and the speed variance is large, it is determined that there is a traffic flow conflict risk. These risk types are abstractions and identifiers of unfavorable conditions in the physical world. They can exist individually, but more often multiple types coexist.

[0033] Furthermore, it is necessary to determine the risk distribution type. When identifying risk distribution types, the focus is on the geometric shape of risk characteristics in the road network space. Using the high-precision spatial coordinates (such as road station numbers) carried by each feature data point in the first fused data, cluster analysis and continuity assessment are performed on the spatial locations of risk feature points of the same type. Specifically, the clustering of feature points is identified through a spatial density algorithm. If high-risk feature points, such as low-friction coefficient points, appear densely only in a very small local area (e.g., a bridge surface or a curve), while the feature values ​​of the surrounding area are normal, their spatial distribution is identified as point-like. If high-risk feature points appear continuously along the longitudinal direction of the road, forming a significantly long and coherent strip-like distribution, they are identified as segment-like. If high-risk features are uniformly or universally present in a large area, almost covering the entire monitored road segment, they are identified as region-like. This identification process essentially adds key geometric attributes to the risk type, clarifying the breadth and spatial pattern of the risk's impact.

[0034] The identified risk types are combined with the identified risk distribution types through combination and weight analysis to output concise and clear scenario labels. It is understandable that different combinations of types and distributions correspond to different safety hazards and response strategies. For example, the combination of low visibility risk and regional distribution identifies regional visibility risk as the dominant risk scenario, and the focus of response is on global speed control and contour enhancement; while the combination of slippery road surface risk and point distribution identifies point icing risk as the dominant risk scenario, and the core of response is focused warnings and upstream early warnings for localized points. When multiple risk types coexist, based on preset rules, such as prioritizing the risk most likely to cause serious accidents, the dominant risk type is calculated and then combined with the distribution type to form a unified and unambiguous scenario instruction. This provides a precise index for subsequently calling highly matched integrated control solutions from the collaborative strategy rule base.

[0035] Furthermore, the data collected by the multi-source sensors at the roadside end also includes vehicle trajectory data. The multi-source sensors also include multi-source sensors at the vehicle end for collecting vehicle-side data; the vehicle-side data includes driver vehicle operation data.

[0036] The vehicle-road cooperative module also includes a driving behavior analysis unit, which is used to calculate a dynamic behavior index reflecting the driver's state based on the vehicle trajectory data and vehicle operation data.

[0037] Furthermore, the instruction generation unit is also used to identify vehicles in high-risk driving states based on the dynamic behavior index, and to adjust the intensity and content of the instructions issued to vehicles in high-risk driving states.

[0038] In this embodiment of the invention, vehicle trajectory data primarily originates from the continuous observation and calculation of the external motion state of vehicles on the road by roadside sensing units (such as high-definition cameras and millimeter-wave radar). Through target detection and multi-target tracking algorithms based on image sequences, the continuous positions of vehicles in the image coordinate system are obtained, and then converted to the world coordinate system using camera calibration parameters to form a position trajectory. Simultaneously, high-precision position, velocity vector, and acceleration information of vehicles are obtained from the raw point cloud data of the millimeter-wave radar through clustering, tracking filtering, and other processing. After spatiotemporal alignment and correlation verification, this information is aggregated into structured vehicle trajectory data, the core content of which includes at least each vehicle's unique temporary identifier, timestamp, three-dimensional spatial coordinates, instantaneous velocity, instantaneous acceleration, and heading angle, thus comprehensively describing the macroscopic kinematic state of the vehicle in the road network.

[0039] Vehicle operation data aims to reflect the driver's real-time control input to the vehicle. It is primarily acquired through standardized data actively reported by the vehicle itself via vehicle-to-everything (V2X) communication technology. This data originates from the vehicle's internal controller area network and onboard sensors, but is provided externally in an anonymized and aggregated form that does not involve personally identifiable information. Specific data may include, but is not limited to: steering wheel angle and rate of turn, used to determine steering smoothness and urgency; accelerator and brake pedal opening and rate of change, reflecting the intensity of acceleration and deceleration intentions; the on / off status of turn signals and hazard warning lights, reflecting the driver's communication of intent and compliance with regulations; and indirect signals triggered by the electronic stability control system or anti-lock braking system.

[0040] Finally, a dynamic behavior index reflecting the driver's state is calculated based on the two types of data mentioned above. This index aims to assess the unconventionality or risk tendency of the driver's behavior in real time. Specifically, firstly, a set of spatiotemporal features reflecting the abnormality of vehicle movement is extracted from the real-time incoming vehicle trajectory data, such as the standard deviation of lateral position, the variance of longitudinal acceleration, and the time distance fluctuation with respect to the vehicle ahead. Simultaneously, another set of features reflecting the aggressiveness of handling is extracted from the vehicle operation data, such as the peak value of steering wheel angle rate and the frequency of rapid acceleration and deceleration events. Subsequently, these heterogeneous features are input into a trained evaluation function, which can characterize the deviation between safe driving behavior features and the currently extracted features under stressful environments such as complex weather conditions. Finally, the dynamic behavior index is output. The higher the index, the greater the deviation of the current driving behavior from the safety benchmark model, that is, the more unstable or aggressive the driving state.

[0041] The core of personalized early warning lies in identifying vehicles with high-risk driving conditions based on the dynamic behavior index and implementing corresponding instructions. Specifically, a short-term historical window of the index is maintained for each trackable vehicle in real-time, with a threshold set. When a vehicle's dynamic behavior index consistently exceeds the threshold, it is marked as a high-risk driving vehicle. For such vehicles, the instruction generation unit adjusts the intensity and content of the general collaborative control instructions. For example, in terms of intensity adjustment, higher-priority, stronger warning voice or text messages are sent to the vehicle's onboard terminal via vehicle-to-infrastructure communication; simultaneously, the roadside rear-end collision warning device can be instructed to trigger a brighter or uniquely frequencyd directional light strip flashing when the vehicle approaches, enhancing its visual appeal. In terms of content adjustment, in addition to the general "slippery road ahead" information, more targeted content may be added, such as "You are following too closely, please increase your distance" or "Please drive smoothly," providing behavioral correction prompts. By performing the above operations, limited security alert resources are allocated in a differentiated manner, providing the most comprehensive and effective intervention to those who need the alerts the most, thereby maximizing security effectiveness at the system level.

[0042] Furthermore, the evaluation function used to calculate the dynamic behavior index is constructed and trained to integrate multidimensional and heterogeneous vehicle motion and operation characteristics into a scalar value that can quantify the driver's real-time risk status. Preferably, the evaluation function is implemented as a rule-weighted linear scoring model. Its construction process first uses traffic psychology and vehicle dynamics principles to select a set of features with clear safety indications from vehicle trajectory data and vehicle operation data as model input. This input feature set includes, but is not limited to: features reflecting lateral control stability, such as the standard deviation of the vehicle's lateral position relative to the lane centerline; features reflecting longitudinal following behavior stability, such as the variance of longitudinal acceleration and the coefficient of variation of the distance to the front vehicle; and features reflecting the aggressiveness of handling, such as the maximum value within a sliding window of the steering wheel angle rate, and the count of rapid acceleration and deceleration events per unit time (defined as events where the absolute value of acceleration exceeds a set threshold). Each input feature needs to undergo normalization preprocessing before being input into the model, that is, its original value is mapped to the [0,1] interval. The upper and lower limits of the normalization can be determined based on historical statistical data and / or theoretical safety thresholds under a large number of normal driving scenarios.

[0043] The aforementioned evaluation function achieves feature fusion by assigning a weight coefficient to each normalized input feature. The determination of the weight coefficient reflects the contribution of each feature to the overall driving risk state. The weights are initially set using expert experience and further calibrated and optimized using historical driving data (including normal driving and typical dangerous driving segments), for example, by employing multiple linear regression to minimize the error between the model output value and the manually labeled risk level labels based on historical data. The output of this function is a normalized dynamic behavior index ranging from [0,1]. The higher the value, the greater the deviation of the current driving behavior pattern from the safety benchmark, i.e., the higher the probability that the driver is in a high-risk state. As an alternative implementation, the evaluation function can also be constructed using lightweight machine learning models such as gradient boosting decision trees or support vector machines. In this case, the training process requires supervised learning of the model parameters using a historical driving feature dataset labeled with risk levels. The trained model uses the same feature set as input and directly outputs the predicted dynamic behavior index. Regardless of the specific function form used, the goal is to achieve a continuous, objective, and quantitative assessment of the driver's condition, providing a basis for decision-making in subsequent personalized warnings.

[0044] Furthermore, the specific type of vehicle includes at least one of school buses, public buses, large passenger vehicles, trucks, and dangerous goods transport vehicles.

[0045] The vehicle-road cooperative control module also configures different safety parameter models for different specific types of vehicles, and when generating cooperative control commands based on the first fused data, it also includes: Based on specific vehicle identification information and corresponding safety parameter models, differentiated instructions are generated for specific types of vehicles entering high-risk road sections.

[0046] It is understandable that the braking distance, center of gravity height, inertial mass, and potential threat to public safety posed by a school bus full of students or a tanker truck transporting flammable and explosive materials are significantly different from those of an ordinary passenger car. Therefore, applying a uniform safety threshold and control strategy is not only inefficient but may also fail to provide sufficient safety margins for high-risk vehicles or cause unnecessary interference to low-risk vehicles.

[0047] In this embodiment of the invention, the safety parameter model is a set of key parameters bound to vehicle type for risk assessment and command generation. Its core parameters include, but are not limited to: a baseline safe following distance coefficient (calculated based on the typical braking performance and reaction time of the vehicle type), a rollover risk coefficient (related to vehicle center of gravity height, wheelbase, and other stability factors), a curve and crosswind sensitivity coefficient, and a social risk weighting factor (used to reflect the additional level of protection required for vehicles with a large number of occupants or high-risk cargo). For example, a model configured for large passenger buses would set a longer baseline safe following distance and a higher rollover risk coefficient; while a model configured for hazardous materials transport vehicles, in addition to the above parameters, would also assign them an extremely high social risk weighting factor, ensuring that they receive the highest level of attention in any early warning decision.

[0048] The construction and determination process of the aforementioned safety parameter model is mainly based on multi-source prior knowledge, which is predefined and solidified in the rule base before system deployment. This belongs to the system design carried out by those skilled in the art based on vehicle engineering knowledge and safety specifications. The input for model construction mainly comes from three types of data: first, vehicle design specifications and dynamic characteristic data, such as typical mass, wheelbase, and braking system performance parameters of various vehicle models obtained from public information or manufacturers, used to calculate the theoretical minimum safe distance through physical formulas; second, traffic accident statistics, analyzing the accident patterns and causes of specific vehicle models under specific weather and road conditions, used to calibrate risk coefficients; and third, traffic regulations and special transportation safety management regulations, which provide a legal basis for setting parameters such as social risk weighting factors. The construction process involves combining these input data with engineering calculations and expert experience judgment to assign specific parameter values ​​to each specific type of vehicle. For example, the output safety parameter model for a large passenger bus might be specifically represented as: {Benchmark safe following distance coefficient: 1.8, rollover risk coefficient: high, social risk weighting factor: 2.0}, while the model for a small passenger bus might be: {Benchmark safe following distance coefficient: 1.0, rollover risk coefficient: low, social risk weighting factor: 1.0}. The model is integrated into the vehicle-road cooperative control module in the form of a lookup table or configuration file.

[0049] In real-time operation, when the roadside perception unit identifies a specific type of vehicle and determines that it is entering or is in a high-risk road segment (such as an icy curve), the instruction generation unit will perform the following steps: First, based on the vehicle's identification information (such as a "dangerous goods transport vehicle" label), it will retrieve parameters that perfectly match the vehicle from the built-in safety parameter model library. Then, these specific parameters will be substituted into the general warning algorithm. For example, when calculating the dynamic rear-end collision prevention warning threshold for all vehicles on the road segment, for this dangerous goods transport vehicle, the product of its baseline safe following distance coefficient (assumed to be 2.5) and social risk weighting factor (assumed to be 3.0) will be used to amplify and correct the general threshold, thereby obtaining a more conservative (i.e., earlier triggering warning) safe distance threshold for this vehicle. Ultimately, based on this customized threshold-based high-risk assessment, enhanced instructions will be triggered for the vehicle. For example, a traffic warning and guidance unit will broadcast a message to following vehicles: "Danger vehicle ahead, please maintain a safe following distance." Simultaneously, actively illuminated road markings will be coordinated to flash a special red pattern a certain distance behind the lane where the dangerous goods vehicle is located, physically warning subsequent vehicles to stay away. Through this series of differentiated processing based on the safety parameter model, this invention achieves precise allocation of safety resources, maximizing the protection of the highest-risk targets, thereby optimizing the overall efficiency of road safety resource allocation.

[0050] Furthermore, the vehicle-road cooperative control module also includes an energy efficiency management unit; The energy efficiency management unit determines the working power consumption mode of each unit in the instruction execution module at different times based on the first fused data and the operation content to be performed by each unit in the instruction execution module; the working power consumption mode includes sleep mode, duty mode, enhanced mode and full power warning mode.

[0051] Furthermore, the energy efficiency management unit is also used to obtain real-time energy status information of each unit of the instruction execution module; the instruction generation unit is also used to adjust the parameters of the collaborative control instruction according to the energy status information.

[0052] In this embodiment of the invention, the core function of the energy efficiency management unit is to achieve dynamic optimization and refined management of energy consumption of each unit of the instruction execution module under the premise of ensuring safety objectives, through intelligent perception and prediction of the system's operating status. This unit receives first fused data from the data fusion unit, from which it analyzes key variables directly affecting energy consumption, mainly including real-time traffic flow density, ambient light intensity, currently identified dominant risk scenarios, and their warning levels. Simultaneously, it receives the operation content to be performed by each unit of the instruction execution module from the instruction generation unit, clarifying the specific tasks required for each unit to complete within a future time period, such as "displaying the number 60," "performing a slow yellow flash," or "maintaining constant brightness at 70%." Based on these inputs, the energy efficiency management unit first assesses the safety necessity; that is, high-risk scenarios and high warning levels mean that the relevant execution units must operate with sufficient performance (brightness, refresh rate), at which point energy-saving objectives are secondary. Secondly, assess energy-saving potential. For example, on road sections with near-zero traffic flow and good weather late at night, even when under monitoring, the safety effectiveness of most illuminated road markings and warning devices is extremely low, indicating significant potential for energy reduction. Similarly, during the day with ample sunlight, dot-matrix LED signs require extremely high brightness to be clearly visible, while on moonlit nights, their base brightness can be significantly reduced. Finally, based on a pre-defined energy efficiency strategy rule base, assign appropriate operating power consumption modes to each execution unit.

[0053] Through the above operations, the embodiments of the present invention significantly reduce the overall operating energy consumption and maintenance costs of the system without sacrificing or even enhancing active safety capabilities, thereby achieving the feasibility of large-scale deployment and the sustainability of long-term operation.

[0054] Example 2: The second aspect of the present invention discloses a method for early warning of traffic safety in complex weather based on vehicle-road cooperative systems using the Internet of Things, the method comprising: Multi-source data is collected through multi-source sensors; the multi-source sensors include roadside sensors, and the collected data includes road surface condition data, traffic flow data, meteorological data, and specific type vehicle identification data; The multi-source data are fused to generate the first fused data; Generate collaborative control instructions based on the first fused data; The coordinated control commands are executed in concert by the dynamic speed limit display unit, the driving warning and guidance unit, and the active luminous traffic marking unit.

[0055] Furthermore, the process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction.

[0056] It should be noted that the specific implementation process of Embodiment 2 is similar to that of Embodiment 1, and will not be repeated in this embodiment.

[0057] Finally, it should be noted that the above-described embodiments include multiple parallel implementations of the present invention. Deleting or otherwise adjusting one or more implementations will not affect the implementation of the solution. Furthermore, the IoT-based vehicle-road cooperative complex weather traffic safety early warning system and method disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit it. Although the present 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; and 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 the present invention.

Claims

1. An Internet of Things-based vehicle-road cooperation complex weather traffic safety warning system, the system comprising an instruction execution module; the instruction execution module comprises a dynamic speed limit display unit, a driving warning guidance unit and an active light-emitting traffic marking line unit; characterized in that, The system also includes a vehicle-road cooperative control module; wherein, the vehicle-road cooperative control module specifically includes: The sensing unit is used to collect multi-source data through multi-source sensors; the multi-source sensors include sensors at the roadside end, and the collected data includes road surface condition data, traffic flow data, meteorological data, and vehicle identification data of specific types; The data fusion unit is used to receive and fuse the multi-source data to generate first fused data; The instruction generation unit is used to generate collaborative control instructions based on the first fused data, and send the collaborative control instructions to the instruction execution module; The instruction execution module is used to receive the collaborative control instruction and execute the collaborative control instruction in coordination with the dynamic speed limit display unit, the driving warning guidance unit, and the active luminous traffic marking unit. 2.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 1, characterized in that, The instruction generation unit has a built-in collaborative strategy rule library; The process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction. 3.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 2, characterized in that, The process of determining the dominant risk scenario based on the first fused data includes: The risk type is determined based on the first fusion data; Based on the spatial distribution of sensor data, risk distribution types are identified; the risk distribution types include point-like, segment-like, and region-like distributions. The dominant risk scenario is determined based on the risk type and risk distribution type. 4.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 2, characterized in that, The data collected by the multi-source sensors at the roadside end also includes vehicle trajectory data; the multi-source sensors also include multi-source sensors at the vehicle end for collecting vehicle-end data; the vehicle-end data includes driver's vehicle operation data; The vehicle-road cooperative module also includes a driving behavior analysis unit, which is used to calculate a dynamic behavior index reflecting the driver's state based on the vehicle trajectory data and vehicle operation data. 5.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 4, characterized in that, The instruction generation unit is also used to determine high-risk driving state vehicles based on the dynamic behavior index, and to adjust the intensity and content of instructions issued to high-risk driving state vehicles. 6.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 2, characterized in that, The specific type of vehicle includes at least one of school buses, public buses, large passenger vehicles, trucks, and dangerous goods transport vehicles; The vehicle-road cooperative control module also configures different safety parameter models for different specific types of vehicles, and when generating cooperative control commands based on the first fused data, it also includes: Based on specific vehicle identification information and corresponding safety parameter models, differentiated instructions are generated for specific types of vehicles entering high-risk road sections. 7.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to any one of claims 1-6, characterized in that, The vehicle-road cooperative control module also includes an energy efficiency management unit; The energy efficiency management unit determines the working power consumption mode of each unit in the instruction execution module at different times based on the first fused data and the operation content to be performed by each unit in the instruction execution module; the working power consumption mode includes sleep mode, duty mode, enhanced mode and full power warning mode. 8.The Internet of Things based cooperative vehicle infrastructure complex weather traffic safety warning system according to claim 7, characterized in that, The energy efficiency management unit is also used to obtain real-time energy status information of each unit of the instruction execution module; the instruction generation unit is also used to adjust the parameters of the collaborative control instruction according to the energy status information. 9.A method for complex weather traffic safety warning based on Internet of Things, the method is applied to the system of any one of claims 1-8, characterized in that, The method includes: Multi-source data is collected through multi-source sensors; the multi-source sensors include roadside sensors, and the collected data includes road surface condition data, traffic flow data, meteorological data, and specific type vehicle identification data; The multi-source data are fused to generate the first fused data; Generate collaborative control instructions based on the first fused data; The coordinated control commands are executed in concert by the dynamic speed limit display unit, the driving warning and guidance unit, and the active luminous traffic marking unit. 10.The IoT-based complex weather traffic safety warning method of CVC according to claim 9, wherein, The process of generating collaborative control instructions based on the first fused data specifically includes: The dominant risk scenario is determined based on the first fusion data, and a preset collaborative control mode is determined based on the dominant risk scenario and the collaborative strategy rule base. The collaborative control mode defines the priority and parameter combination relationship among the three collaborative control instructions: dynamic speed limit instruction, rear-end collision prevention risk level instruction, and active luminous road marking control instruction.

Images