Traffic safety early warning method and equipment for dangerous cargo transport vehicle

Abstract

The invention discloses a traffic safety early warning method and device for a dangerous cargo transport vehicle, and the method comprises the steps: S1, monitoring the operation state of the dangerous cargo transport vehicle in real time, and collecting environment data in real time; and S2, based on the collected data, establishing a risk assessment model by using a convolutional neural network algorithm, analyzing the influence of extreme weather and complex terrains on the safety of the dangerous cargo transport vehicle, developing a risk assessment algorithm, and according to a risk assessment result, the risk assessment model established by using the convolutional neural network algorithm in the invention can be used for evaluating the safety of the dangerous cargo transport vehicle. And specific influences of extreme weather and complex terrains on vehicle safety can be deeply analyzed. The risk assessment model can learn various risk factors in a complex environment and comprehensively assess the risk factors, so that the accuracy and reliability of risk assessment are improved. The system can continuously monitor the change condition of the risk, timely adjust the early warning strategy according to the actual condition, and continuously improve the adaptive capacity and the early warning effect.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic safety control, and in particular to a traffic safety early warning method and device for vehicles transporting dangerous goods. Background Art

[0002] Dangerous goods refer to substances and articles that possess hazardous properties such as explosion, flammability, toxicity, infection, corrosion, and radioactivity, and that can easily cause personal injury, property damage, or environmental pollution during transportation, storage, production, operation, use, and disposal, requiring special protection. These goods can be categorized by their hazardous properties as explosives, compressed and liquefied gases, flammable liquids, flammable solids, spontaneously combustible materials and materials that easily ignite when wet, oxidizers and organic peroxides, toxic and infectious materials, radioactive materials, and corrosive materials.

[0003] A search revealed Chinese patent number CN118212766A, which discloses a traffic safety warning method and device for dangerous goods transport vehicles. The method comprises: obtaining traffic status information on the main road where the dangerous goods transport vehicle is located; the traffic status information includes at least information on pedestrian status in the waiting area, vehicle status on the main road, environmental status, traffic control equipment status, and dangerous goods transport vehicle status; generating a comprehensive decision-making strategy based on the traffic status information; and controlling the traffic control equipment on the main road where the dangerous goods transport vehicle is located based on the comprehensive decision-making strategy, and issuing safety warnings for the dangerous goods transport vehicle. By monitoring traffic status information on the main road where the dangerous goods transport vehicle is located, adaptively adjusting the status of the traffic control equipment based on the current traffic status information, and issuing safety warnings for the dangerous goods transport vehicle, the method improves road utilization efficiency and enhances the safety of dangerous goods transport vehicles and pedestrians. Compared to existing technologies, this invention patent number CN118212766A addresses the difficulty of providing comprehensive and systematic warning measures for the traffic safety of dangerous goods transport vehicles while also ensuring road traffic efficiency due to the long intervals between traffic control equipment and the single source of information.

[0004] However, during the actual use of the above invention, the system's adaptability may be insufficient in extreme weather or complex terrain, making it difficult to accurately assess risks. Due to factors such as slippery roads, low visibility, steep slopes, and sharp turns, vehicles transporting dangerous goods are more prone to accidents such as rollovers, rear-end collisions, and leakage, which may pose a huge threat to people's lives, property, and the environment. Therefore, a traffic safety warning method and equipment for vehicles transporting dangerous goods are proposed. Summary of the Invention

[0005] The purpose of this invention is to solve the shortcomings of the existing technology, that is, the system's adaptability may be insufficient in extreme weather or complex terrain, making it difficult to accurately assess risks. Due to factors such as slippery roads, low visibility, steep slopes, and sharp turns, dangerous goods transport vehicles are more prone to accidents such as rollovers, rear-end collisions, and leakage, which may pose a huge threat to people's lives, property safety, and the environment. The proposed traffic safety warning method and equipment for dangerous goods transport vehicles are intended to address the shortcomings of the existing technology, that is, the system's adaptability may be insufficient in extreme weather or complex terrain, making it difficult to accurately assess risks. Due to factors such as slippery roads, low visibility, steep slopes, and sharp turns, dangerous goods transport vehicles are more prone to accidents such as rollovers, rear-end collisions, and leakage, which may pose a huge threat to people's lives, property safety, and the environment.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] Traffic safety warning method for dangerous goods transport vehicles, including:

[0008] S1: Real-time monitoring of the operating status of dangerous goods transport vehicles and real-time collection of environmental data;

[0009] S2: Based on the collected data, a convolutional neural network algorithm is used to establish a risk assessment model to analyze the impact of extreme weather and complex terrain on the safety of dangerous goods transport vehicles. A risk assessment algorithm is developed. Based on the risk assessment results, the risks are classified into different levels and corresponding early warning strategies are formulated.

[0010] S3: When potential risks are detected, the system automatically generates warning information and conveys it to the driver and relevant management personnel in a timely and accurate manner;

[0011] S4: Drivers and managers initiate corresponding emergency plans based on the received warning information;

[0012] S5: The system continuously monitors changes in risks, adjusts early warning strategies in a timely manner, and collects feedback for improvement and optimization.

[0013] The above technical solution further includes:

[0014] Preferably, the vehicle operating status in S1 includes key parameters such as vehicle speed, position, braking system, and tire condition; the environmental data includes a weather module and a terrain module; the weather module includes temperature, humidity, wind speed, and precipitation information; and the terrain module includes slope, curvature, road condition, historical accident data, and traffic flow data.

[0015] Preferably, the data fusion technology is used in S2 to fuse the collected different data. The specific steps of the data fusion technology are:

[0016] Data collection: collect environmental data from each sensor in real time;

[0017] Time synchronization: synchronize the data from different sensors to ensure that all data are fused and processed at the same timestamp;

[0018] Preprocessing and feature extraction: Model selection based on the needs of the application scenario;

[0019] Data fusion algorithm: Select Kalman filter algorithm or particle filter algorithm to fuse data from different sensors;

[0020] Result evaluation: Evaluate the fusion results to ensure that they can effectively reflect the actual environment;

[0021] Feedback and optimization: Adjust and optimize the data fusion process based on evaluation results and actual application feedback.

[0022] Preferably, the specific steps of establishing the risk assessment model using the convolutional neural network algorithm in S2 include:

[0023] Data preparation: Collect driving data of dangerous goods transport vehicles in extreme weather and complex terrain, preprocess and standardize the collected data, and divide the data into training, validation, and test sets;

[0024] Model construction: The input layer receives multidimensional input data and converts it into a form that the model can process. The convolution layer extracts features and uses multiple convolution kernels for feature learning. After the convolution operation, ReLU is applied to perform nonlinear transformations on the feature map to increase the network's expressive power. By selecting the average value within the local perception area as the pooling result, the spatial dimension of the feature map is reduced while retaining important feature information. After the convolution and pooling layers, a fully connected layer is constructed to integrate features and output risk assessment results. The cross-entropy loss function is used to measure the difference between the model output and the actual risk.

[0025] Model training and validation: Use the training set to train the model, adjust the model parameters through iterative optimization algorithms, and use the validation set to validate the model and evaluate the model performance;

[0026] Model testing and optimization: Use the test set to test the model, evaluate the model's performance, and optimize the model based on the test results.

[0027] Preferably, the risk assessment model regularly updates the environmental status according to the newly collected data.

[0028] E(t)=E(t-1)+ΔE

[0029] Where E(t) represents the environmental state at time t, and ΔE represents the environmental change within the time interval.

[0030] Preferably, the formula of the risk assessment algorithm in S2 is

[0031]

[0032] Where R represents the total risk score, ω i represents the weight of the i-th risk factor, f i (D) represents the function of the i-th risk factor, calculated based on the environmental data D.

[0033] Preferably, the formula for dividing risks into different levels in S2 is:

[0034]

[0035] Among them, R1 and R2 represent the thresholds of the warning level

[0036] Preferably, the emergency plan in S4 includes providing an optimal driving route, an optimal driving speed, maintaining a safe distance, and finding a safe area to park.

[0037] Preferably, the step of collecting feedback in S5 includes:

[0038] Establish a dedicated feedback platform: Establish a dedicated feedback platform to facilitate drivers and managers to submit opinions and suggestions anytime and anywhere;

[0039] Regular face-to-face communication: regularly organize seminars for drivers and managers;

[0040] Use questionnaires: Design detailed questionnaires covering various aspects of early warning methods and send them to drivers and managers for completion.

[0041] Traffic safety warning equipment for vehicles transporting dangerous goods includes: meteorological sensors, GPS, lidar, and cameras; the meteorological sensors are used to monitor weather conditions in real time, the GPS is used to provide precise location and speed information of the vehicle, the lidar is used to obtain high-precision three-dimensional data of the surrounding environment, and the cameras are used to capture real-time images and identify traffic signs, pedestrians, and other vehicles.

[0042] The present invention has the following beneficial effects:

[0043] 1. The risk assessment model established in this invention, using a convolutional neural network algorithm, can deeply analyze the specific impacts of extreme weather and complex terrain on vehicle safety. This risk assessment model can learn from multiple risk factors in complex environments and conduct a comprehensive assessment, thereby improving the accuracy and reliability of risk assessments. The system can continuously monitor changes in risk and promptly adjust early warning strategies based on actual conditions, continuously improving its adaptability and effectiveness.

[0044] 2. In this invention, the system can classify risks into different levels based on risk assessment results and develop corresponding early warning strategies. This helps drivers and managers take appropriate countermeasures based on the severity of the risk, effectively preventing accidents. When the system detects potential risks, it automatically generates early warning information and communicates it to drivers and relevant management personnel in a timely and accurate manner. This timely early warning mechanism enables relevant personnel to respond immediately, avoiding or mitigating the consequences of accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 This is a system architecture diagram of the traffic safety warning method for dangerous goods transport vehicles proposed in the present invention. DETAILED DESCRIPTION

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] like Figure 1 As shown, the traffic safety warning method for dangerous goods transport vehicles proposed by the present invention includes:

[0048] S1: Real-time monitoring of the operating status of dangerous goods transport vehicles and real-time collection of environmental data;

[0049] S2: Based on the collected data, a convolutional neural network algorithm is used to establish a risk assessment model to analyze the impact of extreme weather and complex terrain on the safety of dangerous goods transport vehicles. A risk assessment algorithm is developed. Based on the risk assessment results, the risks are classified into different levels and corresponding early warning strategies are formulated.

[0050] S3: When potential risks are detected, the system automatically generates warning information and conveys it to the driver and relevant management personnel in a timely and accurate manner;

[0051] S4: Drivers and managers initiate corresponding emergency plans based on the received warning information;

[0052] S5: The system continuously monitors changes in risks, adjusts early warning strategies in a timely manner, and collects feedback for improvement and optimization.

[0053] In an embodiment of the present invention, the method first obtains the operating status and environmental data of the dangerous goods transport vehicle, and transmits these data to the central server via a wireless network to provide a basis for subsequent analysis. Based on the rich data collected, an efficient risk assessment model is constructed through a neural network algorithm. The model can deeply learn and understand the specific impact mechanisms of extreme weather such as heavy rain, strong winds, lightning, ice and snow, and complex terrain such as steep mountain roads, sharp bends, tunnels, bridges, etc. on the safety of dangerous goods transport vehicles. Through continuous training and optimization, the model can accurately predict the types and levels of risks that vehicles may face under different weather and terrain conditions. This method improves the adaptability of the system in extreme weather or complex terrain, facilitates accurate risk assessment, and prevents accidents such as rollovers, rear-end collisions, and leakage of dangerous goods transport vehicles.

[0054] In one embodiment, the vehicle operating status in S1 includes key parameters such as vehicle speed, position, braking system, and tire condition; the environmental data includes a weather module and a terrain module; the weather module includes temperature, humidity, wind speed, and precipitation information; and the terrain module includes slope, curvature, road condition, historical accident data, and traffic flow data.

[0055] In one embodiment, the data fusion technology is used in S2 to fuse the collected different data. The specific steps of the data fusion technology are as follows:

[0056] Data collection: collect environmental data from each sensor in real time;

[0057] Time synchronization: synchronize the data from different sensors to ensure that all data are fused and processed at the same timestamp;

[0058] Preprocessing and feature extraction: Model selection based on the needs of the application scenario;

[0059] Data fusion algorithm: Select Kalman filter algorithm or particle filter algorithm to fuse data from different sensors;

[0060] Result evaluation: Evaluate the fusion results to ensure that they can effectively reflect the actual environment;

[0061] Feedback and optimization: According to the evaluation results and actual application feedback, adjust and optimize the data fusion process. In the embodiment of the present invention, the Kalman filter algorithm is used for state estimation and noise elimination, and its formula is:

[0062] x k =F k x k-1 +B k u k +w k

[0063] zk =H K x k +v k

[0064] Among them, x k represents the state vector, F k represents the state transfer matrix, B k represents the control input model, u k represents the control vector, w k and v k Represents process noise and measurement noise. The particle filter algorithm can effectively integrate data from different sensors, improve the accuracy of state estimation, and adapt to complex environments. The specific steps of the particle filter algorithm are:

[0065] Initialization: Generate N particles and its weight

[0066] Prediction step: Make predictions for each particle based on the system dynamic model:

[0067]

[0068] Among them, f is the state transfer function, u k is the control input;

[0069] Update step: Update the weight of each particle using sensor measurements:

[0070]

[0071] Among them, z k Indicates the current measurement value. is the observation model;

[0072] Resampling: Resample particles according to their weights, retaining high-weight particles and discarding low-weight particles.

[0073] In one embodiment, the specific steps of establishing the risk assessment model using the convolutional neural network algorithm in S2 include:

[0074] Data preparation: Collect driving data of dangerous goods transport vehicles in extreme weather and complex terrain, preprocess and standardize the collected data, and divide the data into training, validation, and test sets;

[0075] Model construction: The input layer receives multidimensional input data and converts it into a form that the model can process. The convolution layer extracts features and uses multiple convolution kernels for feature learning. After the convolution operation, ReLU is applied to perform nonlinear transformations on the feature map to increase the network's expressive power. By selecting the average value within the local perception area as the pooling result, the spatial dimension of the feature map is reduced while retaining important feature information. After the convolution and pooling layers, a fully connected layer is constructed to integrate features and output risk assessment results. The cross-entropy loss function is used to measure the difference between the model output and the actual risk.

[0076] Model training and validation: Use the training set to train the model, adjust the model parameters through iterative optimization algorithms, and use the validation set to validate the model and evaluate the model performance;

[0077] Model testing and optimization: Use the test set to test the model, evaluate the model's performance, and optimize the model based on the test results.

[0078] In an embodiment of the present invention, the collection covers driving data of dangerous goods transport vehicles in extreme weather including heavy rain, heavy snow, high temperature, etc., and complex terrain including mountain roads, deserts, city streets, etc. These data include vehicle speed, position, acceleration, direction, meteorological data such as temperature, humidity, wind speed, geographic information such as slope, road conditions, and parameters related to traffic conditions. The collected data is standardized using Z-score normalization, and the preprocessed data is divided into training set, validation set, and test set. Typically, the training set accounts for about 70%, the validation set accounts for 15%, and the test set accounts for 15%. This division helps in training, tuning, and final evaluation of the model. The formula for Z-score normalization is:

[0079]

[0080] Where Z is the normalized value, X is the original value, u is the mean of the data, and σ is the standard deviation of the data.

[0081] In one embodiment, the risk assessment model periodically updates the environmental status based on newly collected data.

[0082] E(t)=E(t-1)+ΔE

[0083] Where E(t) represents the environmental state at time t, and ΔE represents the environmental change within the time interval.

[0084] In one embodiment, the formula of the risk assessment algorithm in S2 is:

[0085]

[0086] Where R represents the total risk score, ωi represents the weight of the i-th risk factor, f i (D) represents the function of the i-th risk factor, calculated based on the environmental data D.

[0087] In one embodiment, the formula for classifying risks into different levels in S2 is:

[0088]

[0089] Among them, R1 and R2 represent the thresholds of warning levels.

[0090] In an embodiment of the present invention, the threshold of the warning level is determined by collecting historical accident data, environmental data, and vehicle operating status data, including information related to extreme weather and complex terrain. A preliminary analysis is performed on these data to identify the key factors that lead to safety hazards. A risk assessment model established using a convolutional neural network outputs risk scores under different conditions through the model, calculates the distribution of risk scores, understands statistical information such as the minimum, maximum, mean, and standard deviation of the scores, and uses the deciles of the data to help divide the risk levels. Combined with the opinions of industry experts, the division of risk scores is evaluated to ensure that the threshold of the warning level is set reasonably. During the operation of the system, the effectiveness of the threshold is continuously monitored and verified, and dynamically adjusted according to the actual accident situation.

[0091] In one embodiment, the emergency plan in S4 includes providing an optimal driving route, an optimal driving speed, maintaining a safe distance, and finding a safe area to park.

[0092] In the embodiment of the present invention, the optimal driving path is calculated by the Dijkstra algorithm and can be expressed by the formula d(v)=min{d(v),d(u)+ω(u,v)}, where d(v) represents the shortest path length to node v, and ω(u,v) represents the weight of the edge.

[0093] In one embodiment, the step of collecting feedback in S5 includes:

[0094] Establish a dedicated feedback platform: Establish a dedicated feedback platform to facilitate drivers and managers to submit opinions and suggestions anytime and anywhere;

[0095] Regular face-to-face communication: regularly organize seminars for drivers and managers;

[0096] Use questionnaires: Design detailed questionnaires covering various aspects of early warning methods and send them to drivers and managers for completion.

[0097] In one embodiment, a traffic safety warning device for a dangerous goods transport vehicle includes: a meteorological sensor, a GPS, a lidar, and a camera; the meteorological sensor is used to monitor weather conditions in real time, the GPS is used to provide precise location and speed information of the vehicle, the lidar is used to obtain high-precision three-dimensional data of the surrounding environment, and the camera is used to capture real-time images and identify traffic signs, pedestrians, and other vehicles.

[0098] In this embodiment of the present invention, meteorological sensors monitor various weather conditions in real time, such as temperature, humidity, wind speed, precipitation, and visibility. By integrating these sensors, the system can promptly identify the impact of severe weather on driving safety. In the event of extreme weather, the device can automatically generate an alert and notify the driver via the dashboard or mobile app. GPS provides real-time vehicle location, speed, and direction, ensuring accurate tracking of the vehicle's status at all times. Combined with a map navigation system, the system can analyze the driving route in real time, identify potential danger zones, and provide optimization recommendations. LiDAR captures high-precision three-dimensional data of the surrounding environment, creating a detailed model of objects surrounding the vehicle. LiDAR accurately measures distances to other vehicles, obstacles, and roadside structures. In low-visibility or complex environments, LiDAR provides critical environmental awareness, assisting autonomous driving functions and improving overall driving safety. Cameras capture real-time images and possess image processing capabilities, enabling them to identify traffic signs, pedestrians, other vehicles, and potential obstacles. Using image recognition algorithms, the system analyzes traffic flow, identifies traffic light status, and issues warnings when pedestrians are crossing the road, ensuring timely driver response.

[0099] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A traffic safety warning method for dangerous goods transport vehicles, characterized in that: include: S1: Real-time monitoring of the operating status of dangerous goods transport vehicles and real-time collection of environmental data; S2: Based on the collected data, a convolutional neural network algorithm is used to establish a risk assessment model to analyze the impact of extreme weather and complex terrain on the safety of dangerous goods transport vehicles. A risk assessment algorithm is developed. Based on the risk assessment results, the risks are classified into different levels and corresponding early warning strategies are formulated. S3: When potential risks are detected, the system automatically generates warning information and conveys it to the driver and relevant management personnel in a timely and accurate manner; S4: Drivers and managers initiate corresponding emergency plans based on the received warning information; S5: The system continuously monitors changes in risks, adjusts early warning strategies in a timely manner, and collects feedback for improvement and optimization.

2. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The vehicle operating status in S1 includes key parameters such as vehicle speed, position, braking system, and tire condition. The environmental data includes a weather module and a terrain module. The weather module includes temperature, humidity, wind speed, and precipitation information. The terrain module includes slope, curvature, road condition, historical accident data, and traffic flow data.

3. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: In S2, the collected different data are fused using data fusion technology. The specific steps of the data fusion technology are as follows: Data collection: collect environmental data from each sensor in real time; Time synchronization: synchronize the data from different sensors to ensure that all data are fused and processed at the same timestamp; Preprocessing and feature extraction: Model selection based on the needs of the application scenario; Data fusion algorithm: Select Kalman filter algorithm or particle filter algorithm to fuse data from different sensors; Result evaluation: Evaluate the fusion results to ensure that they can effectively reflect the actual environment; Feedback and optimization: Adjust and optimize the data fusion process based on evaluation results and actual application feedback.

4. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The specific steps of establishing the risk assessment model using the convolutional neural network algorithm in S2 include: Data preparation: Collect driving data of dangerous goods transport vehicles in extreme weather and complex terrain, preprocess and standardize the collected data, and divide the data into training, validation, and test sets; Model construction: The input layer receives multidimensional input data and converts it into a form that the model can process. The convolution layer extracts features and uses multiple convolution kernels for feature learning. After the convolution operation, ReLU is applied to perform nonlinear transformations on the feature map to increase the network's expressive power. By selecting the average value within the local perception area as the pooling result, the spatial dimension of the feature map is reduced while retaining important feature information. After the convolution and pooling layers, a fully connected layer is constructed to integrate features and output risk assessment results. The cross-entropy loss function is used to measure the difference between the model output and the actual risk. Model training and validation: Use the training set to train the model, adjust the model parameters through iterative optimization algorithms, and use the validation set to validate the model and evaluate the model performance; Model testing and optimization: Use the test set to test the model, evaluate the model's performance, and optimize the model based on the test results.

5. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The risk assessment model periodically updates the environmental status based on newly collected data. E(t)=E(t-1)+ΔE Where E(t) represents the environmental state at time t, and ΔE represents the environmental change within the time interval.

6. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The formula of the risk assessment algorithm in S2 is: Where R represents the total risk score, ω i represents the weight of the i-th risk factor, f i (D) represents the function of the i-th risk factor, calculated based on the environmental data D.

7. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The formula for classifying risks into different levels in S2 is: Among them, R1 and R2 represent the thresholds of warning levels.

8. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The emergency plan in S4 includes providing the best driving route, the best driving speed, maintaining the vehicle distance, and finding a safe area to park.

9. The traffic safety warning method for dangerous goods transport vehicles according to claim 1, characterized in that: The step of collecting feedback in S5 includes: Establish a dedicated feedback platform: Establish a dedicated feedback platform to facilitate drivers and managers to submit opinions and suggestions anytime and anywhere; Regular face-to-face communication: regularly organize seminars for drivers and managers; Use questionnaires: Design detailed questionnaires covering various aspects of early warning methods and send them to drivers and managers for completion.

10. The traffic safety warning device for dangerous goods transport vehicles according to claim 1, characterized in that: include: Meteorological sensors, GPS, lidar, and cameras; the meteorological sensors are used to monitor weather conditions in real time, the GPS is used to provide precise location and speed information of the vehicle, the lidar is used to obtain high-precision three-dimensional data of the surrounding environment, and the cameras are used to capture real-time images and identify traffic signs, pedestrians, and other vehicles.

Images