Road environment monitoring terminal based on Internet of Things

Abstract

The invention discloses a road environment monitoring terminal based on the Internet of Things, and relates to the technical field of traffic environment monitoring. The road environment monitoring terminal based on the Internet of Things comprises an environment monitoring module, an edge computing module, an Internet of Things communication module, a cloud analysis module, an intelligent linkage module, a power supply module and an equipment self-maintenance module. According to the invention, real-time preprocessing and intelligent analysis are carried out on the environmental data through the edge calculation module, abnormal modes in the environmental data are rapidly identified, and a real-time analysis result is generated, so that the data transmission bandwidth demand and the processing delay are greatly reduced; the cloud analysis module combines historical data and real-time data, and provides a global evaluation result based on a deep learning model, thereby supporting the intelligence of traffic management decisions.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic environment monitoring, and particularly to a highway environment monitoring terminal based on the Internet of Things. Background Art

[0002] The existing highway environment monitoring technology has the following problems: insufficient real-time monitoring ability. Traditional environment monitoring systems mostly use single sensors to collect environmental data, and are unable to comprehensively and real-time dynamically monitor the complex meteorological conditions, air quality and road conditions along the highway. Especially in the event of sudden environmental changes (such as heavy rainfall, road icing, air pollution incidents), the response speed is relatively slow. Low data processing efficiency. Existing monitoring systems generally rely on a centralized data processing architecture, and all data need to be uploaded to the cloud for analysis before results can be obtained. This architecture is prone to delays in the event of network interruption or a sharp increase in data volume, and cannot meet the requirements of highway environment monitoring for real-time and high efficiency. Insufficient intelligent level. Traditional monitoring systems lack intelligent analysis and prediction capabilities, and are unable to combine real-time and historical data for trend prediction or anomaly detection. At the same time, the response mechanism for emergencies mostly relies on manual intervention, and it is difficult to trigger early warnings or link relevant traffic management devices in a timely manner. Insufficient system reliability. In a complex highway environment, the anti-interference and stability of the devices in the existing system are limited, and the fault monitoring and maintenance capabilities for sensors and communication modules are weak. Once a device fails, it will have a greater impact on the overall monitoring function.

[0003] Based on the above problems, there is an urgent need for a technical solution that can monitor the highway environment in real-time, comprehensively and intelligently. Summary of the Invention

[0004] The present invention provides a highway environment monitoring terminal based on the Internet of Things. Through the organic combination of a multi-sensor array, edge computing, cloud analysis and an intelligent linkage module, it can not only monitor the highway environment in real-time, but also perform intelligent analysis and linkage control on the data, thereby providing optimized decision-making support and emergency response capabilities for traffic management departments.

[0005] The highway environment monitoring terminal based on the Internet of Things includes: An environment monitoring module for collecting real-time environmental data along the highway through a sensor array, including meteorological parameters, air quality parameters and road condition parameters. The environmental data is processed into structured information and transmitted to the edge computing module; An edge computing module, connected to the environmental monitoring module, performs real-time processing on environmental data based on an embedded real-time analysis model. The real-time analysis model includes an anomaly detection sub-model and a predictive analysis sub-model. By performing preliminary analysis on environmental data, it generates real-time analysis results, including data anomaly markers, trend prediction results, and event response information, and transmits the real-time analysis results to the Internet of Things communication module and the intelligent linkage module. An Internet of Things communication module, connected to the edge computing module, is used to upload environmental data and real-time analysis results to the cloud analysis module and receive the global evaluation results returned by the cloud analysis module, realizing interconnection and interoperability between terminal devices and stable communication with the cloud platform. A cloud analysis module, connected to the edge computing module through the Internet of Things communication module, performs global analysis and processing on environmental data and real-time analysis results based on a global evaluation model in the cloud. The global evaluation model combines historical data and real-time data to generate global evaluation results, including environmental prediction information, optimization suggestions for traffic management, and response strategies for emergencies, and transmits the global evaluation results back to the Internet of Things communication module. An intelligent linkage module, respectively connected to the edge computing module and the Internet of Things communication module, is used to receive real-time analysis results and global evaluation results and execute corresponding linkage control operations, including dynamically adjusting the operating parameters of traffic management devices, triggering real-time warning devices, or sending control instructions to the management center.

[0006] As a preferred technical solution of the present invention, the environmental monitoring module includes: The sensor array includes: Meteorological sensors, which collect data related to the meteorological conditions along the highway. The meteorological sensors include temperature sensors, humidity sensors, and wind speed sensors, which are used to detect environmental temperature, humidity, and wind speed respectively. Air quality sensors, which detect the pollutant concentration in the air along the highway. The air quality sensors include PM2.5 sensors, PM10 sensors, and carbon dioxide sensors, which are used to detect fine particulate matter, inhalable particulate matter, and carbon dioxide concentration respectively. Road condition sensors, which detect the condition of the highway surface. The road condition sensors include camera devices, infrared sensors, and humidity sensors. Among them, the camera devices are used to collect image data of the highway surface to analyze water accumulation, icing, obstacles, and traffic flow; the infrared sensors are used to detect the temperature distribution of the highway surface to determine whether there is icing or temperature anomaly; the humidity sensors are used to detect the pavement humidity level to determine water accumulation or slippery conditions. A data processing unit, connected to the sensor array, is used to receive environmental data collected by various sensors, process the environmental data, remove abnormal data and noise data, and convert the processed data into structured information in a unified format and then transmit it to the edge computing module.

[0007] As a preferred technical solution of the present invention, the edge computing module includes: A data receiving unit for receiving environmental data from the environmental monitoring module; A data preprocessing unit connected to the data receiving unit for processing the received environmental data, including removing duplicate data and repairing lost data; A real-time analysis unit connected to the data preprocessing unit for intelligently analyzing the processed data based on an embedded real-time analysis model. The real-time analysis model includes an anomaly detection sub-model for detecting abnormal patterns or abnormal events in the data, including waterlogging, icing, or excessive air pollutants; a predictive analysis sub-model for predicting the future change trend of environmental parameters based on time series data, including road icing risk prediction and meteorological condition change trend prediction; A result generation unit connected to the real-time analysis unit for generating real-time analysis results and transmitting the generated real-time analysis results to the Internet of Things communication module and the intelligent linkage module.

[0008] As a preferred technical solution of the present invention, the real-time analysis model includes: An anomaly detection sub-model that uses an anomaly classification algorithm based on support vector machines, with the input environmental data as a feature vector to construct a separating hyperplane in a high-dimensional space for determining whether the data is an abnormal point. The hyperplane is defined by the following formula: , where is the weight vector, is the bias term. If , it is determined as normal data; otherwise, it is determined as abnormal data; The anomaly detection sub-model is trained by extracting feature vectors from the historical environmental dataset, with the classification accuracy of abnormal data and normal data as the optimization objective; A trend prediction sub-model that uses a time series prediction algorithm based on long short-term memory networks, with the time series of environmental data as the input to predict environmental parameters at future time points; The trend prediction sub-model is trained through supervised learning, using historical time series data and their corresponding future target values as training samples, and minimizing the mean squared error as the loss function for optimization; The real-time analysis model is obtained through pre-training. The training data is sourced from historical environmental monitoring data, including meteorological data, air quality data, and road condition data. The training process is completed through a distributed computing platform, including four stages: data preprocessing, feature extraction, model training, and parameter optimization. The cross-validation method is used to verify the model performance, with classification accuracy, prediction error, and model convergence as the model evaluation criteria.

[0009] As a preferred technical solution of the present invention, the Internet of Things communication module includes: A communication protocol unit, used to support wireless communication protocols, including LoRa, NB-IoT, 4G, and 5G communication protocols, and automatically switch the communication mode according to the application scenario, where the application scenario includes low power consumption, high bandwidth, and long-distance transmission; A data transmission unit, connected to the communication protocol unit, used to upload the environmental data and real-time analysis results generated by the edge computing module to the cloud analysis module, and at the same time receive the global evaluation results returned by the cloud analysis module; the data transmission unit encrypts the transmitted data; A data cache unit, connected to the data transmission unit, used to temporarily store environmental data and real-time analysis results when the communication is interrupted or the network is unstable, and automatically upload them after the communication is restored; A communication status monitoring unit, used to monitor the operating status of the communication module in real time, including network connection status, signal strength, and data transmission rate, and send fault information to the device self-maintenance module when the communication is abnormal; A low-power unit, used to optimize the energy consumption of the communication module, and dynamically adjust the communication power and working mode according to the data transmission task.

[0010] As a preferred technical solution of the present invention, the cloud analysis module includes: A data receiving unit, used to receive the environmental data and real-time analysis results uploaded by the Internet of Things communication module, and classify and store the received data. The data is divided into real-time data and historical data according to the time dimension to support subsequent analysis; A global evaluation unit, connected to the data receiving unit, based on the global evaluation model, globally analyzes and processes the received data, and generates the following three types of results: Prediction information of the environment: Based on the deep learning algorithm, by extracting features from real-time environmental data and combining the temporal variation law of historical environmental data, predict future environmental conditions, including meteorological change trends, air quality distribution, and future changes in road conditions; Optimization suggestions for traffic management: Based on graph convolutional neural network and long short-term memory network, by jointly modeling real-time traffic flow data and historical traffic data, analyze the dynamic change characteristics of traffic flow, and generate optimization suggestions for traffic management, including traffic signal cycle adjustment, diversion route planning, and traffic guidance plans; Response strategies for emergencies: Through real-time environmental data and referring to similar event patterns recorded in historical data, identify current emergencies and generate corresponding response strategies, including real-time warning information, road closure instructions, and warning signal triggering plans; The response strategy generation unit, connected to the global evaluation unit, is used to receive environmental prediction information, traffic management optimization suggestions, and emergency response strategies, and integrate these results into specific executable instructions; The result sending unit, connected to the response strategy generation unit, is used to transmit environmental prediction information, traffic management optimization suggestions, and emergency response strategies to the Internet of Things communication module for use by terminal devices or traffic management centers; The model optimization unit is used to iteratively train deep learning algorithms and traffic flow evaluation algorithms based on graph convolutional neural network and long short-term memory network based on uploaded real-time data and historical data, as well as the actual execution feedback of the global evaluation results, to improve the prediction accuracy and response efficiency of the model output; The cloud storage unit is used to store historical environmental data, global evaluation results, and response strategies, providing long-term data support for the global evaluation unit and the model optimization unit.

[0011] As a preferred technical solution of the present invention, the global evaluation model includes: The environmental prediction sub-model: used to generate prediction information of the environment. The environmental prediction sub-model is constructed based on deep learning algorithms, and its training and use include: Extract historical environmental data from the cloud storage unit as the training data set; use the long short-term memory network as the basic model to capture the temporal changes of environmental data; optimize the model by minimizing the mean square error; Input real-time environmental data into the environmental prediction sub-model, and combine historical data to generate environmental prediction information for the future time period; The traffic flow evaluation sub-model, used to generate optimization suggestions for traffic management. The traffic flow evaluation sub-model is constructed based on a combined algorithm of graph convolutional neural network and long short-term memory network, and its training and use include: Extract the traffic flow in the historical environmental data from the cloud storage unit and combine it with the road network topology as the training dataset; use a graph convolutional neural network to model the spatial features of the road network, use a long short-term memory network to model the temporal features of the traffic flow data, extract the temporal features and combine them with the output of the graph convolutional neural network; combine the results of the graph convolutional neural network and the long short-term memory network to generate a spatio-temporal joint representation of the traffic flow, and optimize the model through the mean square error function; Input the traffic flow and the current road network topology, combine with the historical traffic data, generate the evaluation result of the current traffic flow state and the traffic flow prediction within the future time period; generate the optimization suggestions for traffic management according to the prediction result; An emergency event recognition sub-model, which is used to generate the response strategy for emergency events. The emergency event recognition sub-model is constructed based on the combined algorithm of a convolutional neural network and a temporal convolutional network in deep learning. Its training and use include: Extract the historical emergency event data and the corresponding environmental data from the cloud storage unit as the training dataset; use a convolutional neural network to extract the spatial features of the environmental data and the traffic flow data, and use a temporal convolutional network to model the temporal features of the data; combine the spatial features and the temporal features to construct an emergency event recognition model, and optimize it through the cross-entropy loss function; Input the real-time environmental data and the traffic flow data, generate the emergency event detection result; generate the corresponding response strategy according to the detection result.

[0012] As a preferred technical solution of the present invention, the intelligent linkage module includes: A linkage data receiving unit, which is used to receive the real-time analysis result from the edge computing module and the global evaluation result from the cloud analysis module; A linkage strategy parsing unit, which is connected to the linkage data receiving unit, and is used to parse the global evaluation result and generate the corresponding control instruction according to the parsing result; The parsing process includes: judging the meteorological change trend, air quality change and road state prediction in the environmental prediction information to determine whether to trigger the environmental warning instruction; parsing the traffic signal cycle adjustment, diversion route planning and traffic guidance plan in the traffic management optimization suggestion to generate the corresponding traffic signal device control instruction; parsing the road closure, real-time warning and warning signal triggering in the emergency event response strategy to generate the linkage control instruction; A device control unit, which is connected to the linkage strategy parsing unit, and is used to execute the parsed control instruction; The device control includes: traffic signal device control, dynamically adjusting the traffic signal light cycle, switching traffic signs or activating dynamic warning lights; early warning device control, triggering early warning broadcasts, electronic displays or other warning devices to provide real-time early warning information to vehicles or personnel; A remote control module, used to send linkage control instructions to the traffic management center, supporting remote operation and manual intervention; A status feedback unit, used to monitor the device execution status in real time and transmit the execution status and feedback information to the cloud analysis module and the device self-maintenance module for subsequent optimization and fault handling; A priority scheduling unit, connected to the linkage policy parsing unit, used to schedule multiple linkage control instructions according to priority. When multiple linkage control tasks conflict, select and execute the linkage instruction according to the event type and urgency.

[0013] As a preferred technical solution of the present invention, it further includes: A power module, used to intelligently supply power to the highway environmental monitoring terminal, supporting the low-power operation of each module; The power module includes: A solar power supply unit, used to capture light energy and convert it into electrical energy to provide sustainable energy support for the terminal; A energy storage battery unit, connected to the solar power supply unit, used to store excess electrical energy and supply power to the terminal under insufficient light or no-light conditions; A low-power management unit, used to dynamically adjust the power consumption status of each module, and optimize power distribution by analyzing the operation of the terminal and the energy usage requirements; A backup power supply unit, connected to the energy storage battery unit, used to provide emergency power supply when both the solar power supply and the energy storage battery are unavailable, ensuring the basic operation of the terminal; A power status monitoring unit, used to monitor the operation status of the solar power supply unit, the energy storage battery unit and the backup power supply unit in real time, including the battery level, output power and operation efficiency, and transmit the monitoring results to the device self-maintenance module.

[0014] As a preferred technical solution of the present invention, it further includes: A device self-maintenance module, connected to the environmental monitoring module, the edge computing module, the Internet of Things communication module and the power module in the terminal, used to monitor the operation status of each module, perform automatic device fault detection and diagnosis, and transmit the fault diagnosis report to the intelligent linkage module; The device self-maintenance module includes: A status monitoring unit, used to monitor the operation status of the environmental monitoring module, the edge computing module, the Internet of Things communication module, the power module and the intelligent linkage module in real time, including the working status, data transmission status, power status and device connection status; A fault detection unit, connected to the status monitoring unit, is used to analyze the monitoring data and identify anomalies or faults during the operation of the module, including sensor failure, communication interruption, insufficient power, and module hardware failure; A diagnosis unit, connected to the fault detection unit, is used to generate a fault diagnosis report based on the fault detection results, including the cause of the fault, the scope of influence, and priority assessment; A fault response unit, connected to the diagnosis unit, is used to automatically take corresponding response measures according to the fault diagnosis report, including restarting the module, switching the communication path, enabling the backup power supply, or sending a fault alarm; A fault reporting unit is used to send the fault diagnosis report and response measures to the cloud analysis module or the traffic management center through the Internet of Things communication module; A log recording unit is used to record the device operation log and the fault diagnosis log, including time, location, fault type, and processing result, and store the log in the cloud storage unit.

[0015] The present invention has the following advantages: Through a variety of sensor arrays of the environmental monitoring module, the present invention can collect meteorological parameters, air quality parameters, and road condition parameters along the highway in real time. The monitoring data comprehensively covers a variety of key environmental factors along the highway, providing accurate environmental status information for highway management departments and ensuring traffic safety.

[0016] Through the edge computing module, the present invention performs real-time preprocessing and intelligent analysis on environmental data, quickly identifies abnormal patterns in the environmental data, and generates real-time analysis results, reducing the data transmission bandwidth requirement and processing delay; the cloud analysis module combines historical data and real-time data to provide a global evaluation result based on a deep learning model, thereby supporting the intelligence of traffic management decisions.

[0017] Through the intelligent linkage module, according to the real-time analysis result and the global evaluation result, the present invention dynamically adjusts traffic management devices, triggers warning devices, or sends control instructions to the management center. The automated linkage control ability improves the emergency response efficiency in case of emergencies and reduces the time cost of human intervention; Through the device self-maintenance module, the present invention monitors the operation status of the environmental monitoring module, the edge computing module, the communication module, the power module, etc. in real time, automatically detects device faults, and generates a diagnosis report, thereby ensuring the long-term stable operation of the system; through the fault response mechanism, the system can achieve highly reliable operation in a complex highway environment, and can maintain the normal operation of the monitoring function even in case of device failure or communication interruption.

[0018] The present invention is equipped with a solar power supply unit and an energy storage battery unit. Combining with low-power management technology, it can dynamically adjust power distribution according to energy consumption requirements to ensure the long-term operation of the device in remote or gridless highway environments. This way of using green energy conforms to the concept of energy conservation and environmental protection and reduces the operation and maintenance costs of the system. Brief Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only schematic diagrams of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to the provided drawings. Figure 1 It is a schematic structural diagram of a highway environment monitoring terminal based on the Internet of Things adopted in the embodiment of the present invention. Detailed Embodiments

[0020] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the following will further describe the present invention in detail with reference to the drawings. Obviously, the described embodiments are only partial embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0021] Embodiment. A highway environment monitoring terminal based on the Internet of Things, the structure of which is referred to Figure 1 as shown, includes: An environment monitoring module for collecting real-time environmental data along the highway through a sensor array, including meteorological parameters, air quality parameters, and road condition parameters. The environmental data is processed into structured information and transmitted to the edge computing module; The environment monitoring module includes: The sensor array includes: Meteorological sensors for collecting data related to the meteorological conditions along the highway. The meteorological sensors include a temperature sensor, a humidity sensor, and a wind speed sensor for detecting the ambient temperature, humidity, and wind speed respectively; Air quality sensors for detecting the concentration of pollutants in the air along the highway. The air quality sensors include a PM2.5 sensor, a PM10 sensor, and a carbon dioxide sensor for detecting the concentration of fine particulate matter, inhalable particulate matter, and carbon dioxide respectively; Road condition sensing, which detects the condition of the road surface. The road condition sensing includes a camera device, an infrared sensor, and a humidity sensor. Among them, the camera device is used to collect image data of the road surface to analyze water accumulation, icing, obstacles, and traffic flow; the infrared sensor is used to detect the temperature distribution of the road surface to determine whether there is icing or abnormal temperature; the humidity sensor is used to detect the road surface humidity level to determine water accumulation or slippery conditions; A data processing unit, connected to the sensor array, is used to receive environmental data collected by various sensors, process the environmental data, remove abnormal data and noise data, and convert the processed data into structured information in a unified format and then transmit it to the edge computing module.

[0022] An edge computing module, connected to the environmental monitoring module, performs real-time processing on environmental data based on an embedded real-time analysis model. The real-time analysis model includes an anomaly detection sub-model and a predictive analysis sub-model. By performing preliminary analysis on environmental data, real-time analysis results are generated, including data anomaly markers, trend prediction results, and event response information, and the real-time analysis results are transmitted to the IoT communication module and the intelligent linkage module; The edge computing module includes: A data receiving unit, used to receive environmental data from the environmental monitoring module; A data preprocessing unit, connected to the data receiving unit, is used to process the received environmental data, including removing duplicate data and repairing missing data; A real-time analysis unit, connected to the data preprocessing unit, is used to perform intelligent analysis on the processed data based on an embedded real-time analysis model. The real-time analysis model includes an anomaly detection sub-model, used to detect abnormal patterns or abnormal events in the data, including water accumulation, icing, or excessive air pollutants; a predictive analysis sub-model, used to predict the future change trend of environmental parameters based on time series data, including road icing risk prediction, meteorological condition change trend prediction; A result generation unit, connected to the real-time analysis unit, is used to generate real-time analysis results and transmit the generated real-time analysis results to the IoT communication module and the intelligent linkage module.

[0023] The real-time analysis model includes: An anomaly detection sub-model, which uses an anomaly classification algorithm based on support vector machines, with the input environmental data as a feature vector to construct a separating hyperplane in a high-dimensional space for determining whether the data is an abnormal point. The hyperplane is defined by the following formula: , where is the weight vector, is the bias term. If , it is determined to be normal data, otherwise it is determined to be abnormal data; The abnormal detection sub-model is trained by extracting feature vectors from the historical environmental dataset, with the classification accuracy of abnormal data and normal data as the optimization objective; The trend prediction sub-model uses a time series prediction algorithm based on a long short-term memory network, with the time series of environmental data as the input to predict the environmental parameters at future time points; The core formulas of the long short-term memory network include: Input gate: , where represents the activation value of the input gate (range ), is the weight matrix of the input gate, represents the Sigmod activation function, represents the hidden state of the previous time step, represents the current input, is the bias of the input gate; it is used to determine how much information of the current input is introduced into the cell state .

[0024] Forget gate: , where represents the activation value of the forget gate (range ), is the weight matrix of the forget gate, is the bias of the forget gate; it is used to determine how much information in the state of the previous time step needs to be forgotten; Output gate: , where represents the activation value of the output gate (range ), is the weight matrix of the output gate, is the bias of the output gate; it is used to determine which information in the cell state needs to be output to the current hidden state ; Cell state update: , where is the weight matrix for generating new information, is the current bias term, tanh is the hyperbolic tangent function, which is used to map the information to , representing the importance of the information, represents the state information of the previous time step retained by the forget gate, represents the new information determined by the input gate to be added; it is used to integrate the information of the forget gate and the input gate to update the cell state of the current time step; Hidden state: , represents the hidden state at the current time step, represents the activation value of the output gate, which determines which information is output from the cell state; The trend prediction sub-model is trained through supervised learning. Using historical time series data and their corresponding future target values as training samples, the mean squared error is minimized as the loss function for optimization. The calculation formula for minimizing the mean squared error is: , where represents the true value, is the predicted value, is the total number of samples; The real-time analysis model is obtained through pre-training. The training data comes from historical environmental monitoring data, including meteorological data, air quality data, and road condition data; the training process is completed through a distributed computing platform, including four stages: data preprocessing, feature extraction, model training, and parameter optimization; the cross-validation method is used to verify the model performance, and the classification accuracy, prediction error, and model convergence are used as the model evaluation criteria.

[0025] The specific pre-training process is as follows: Obtain historical environmental monitoring data and perform data preprocessing, including: Data cleaning, removing outliers, noise data, and filling in missing values to ensure the quality of the input data; Data standardization, normalizing the data collected by different sensors to unify the scale range of features; Feature engineering, anomaly detection sub-model: extract key features (such as temperature fluctuation range, humidity change rate, abnormal peak of pollutant concentration, etc.); trend prediction sub-model: generate time series features (data values of the past N time steps); Model training: Anomaly detection sub-model, based on the support vector machine (SVM) algorithm, uses normal samples and abnormal samples in historical data to construct a training set, with feature vectors as input and labels of "normal" or "abnormal", optimizes the classification hyperplane, and optimizes hyperparameters through grid search.

[0026] Trend prediction sub-model, based on the long short-term memory network (LSTM), with time series as input features and environmental parameters at future time points as output targets, uses the Adam optimizer to adjust model parameters to improve the training convergence speed; Model optimization: Cross-validation, dividing the dataset into a training set, a validation set, and a test set; using K-fold cross-validation to evaluate the model performance and optimize the classification accuracy and prediction error; Efficiently train massive historical data based on a distributed computing platform (TensorFlow distributed training framework); parallel computing to accelerate the feature extraction, parameter update, and model verification processes; Regularly add the real-time collected environmental data to the training dataset for model retraining and optimization to ensure the model's adaptability to the latest environmental changes; Performance evaluation metrics of the model. For the anomaly detection sub-model: classification accuracy, recall rate, and F1 score; for the trend prediction sub-model: mean squared error (MSE) and mean absolute error (MAE); After the system is deployed, the real-time collected data will be periodically uploaded to the cloud as the data source for model retraining; through incremental training, maintain the effectiveness of the model during long-term operation; Before deployment: fully train the model using an offline dataset; After deployment: optimize the model according to the actual application scenario through fine-tuning; The IoT communication module, connected to the edge computing module, is used to upload environmental data and real-time analysis results to the cloud analysis module and receive the global evaluation results returned by the cloud analysis module to achieve interconnection between terminal devices and stable communication with the cloud platform; The IoT communication module includes: A communication protocol unit, used to support wireless communication protocols, including LoRa, NB-IoT, 4G, and 5G communication protocols, automatically switch the communication mode according to the application scenario, and the application scenarios include low power consumption, high bandwidth, and long-distance transmission; A data transmission unit, connected to the communication protocol unit, is used to upload the environmental data and real-time analysis results generated by the edge computing module to the cloud analysis module and simultaneously receive the global evaluation results returned by the cloud analysis module; the data transmission unit encrypts the transmitted data; A data caching unit, connected to the data transmission unit, is used to temporarily store environmental data and real-time analysis results when the communication is interrupted or the network is unstable and automatically upload them after the communication is restored; A communication status monitoring unit, used to monitor the operating status of the communication module in real time, including network connection status, signal strength, and data transmission rate, and send fault information to the device self-maintenance module when the communication is abnormal; A low-power unit, used to optimize the energy consumption of the communication module and dynamically adjust the communication power and working mode according to the data transmission task.

[0027] The cloud analysis module is connected to the edge computing module through the Internet of Things communication module, and globally analyzes and processes environmental data and real-time analysis results based on the global evaluation model in the cloud. The global evaluation model combines historical data and real-time data to generate a global evaluation result, including prediction information of the environment, optimization suggestions for traffic management, and response strategies for emergencies, and transmits the global evaluation result back to the Internet of Things communication module; The cloud analysis module includes: A data receiving unit, which is used to receive the environmental data and real-time analysis results uploaded by the Internet of Things communication module, and classify and store the received data. The data is divided into real-time data and historical data according to the time dimension to support subsequent analysis; A global evaluation unit, connected to the data receiving unit, globally analyzes and processes the received data based on the global evaluation model, and generates the following three types of results: Prediction information of the environment: Based on the deep learning algorithm, by extracting features from real-time environmental data and combining the temporal change laws of historical environmental data, predict future environmental conditions, including meteorological change trends, air quality distribution, and future changes in road conditions; Optimization suggestions for traffic management: Based on the graph convolutional neural network and long short-term memory network, through the joint modeling of real-time traffic flow data and historical traffic data, analyze the dynamic change characteristics of traffic flow, and generate optimization suggestions for traffic management, including traffic signal cycle adjustment, diversion route planning, and traffic guidance plans; Response strategies for emergencies: Through real-time environmental data, and referring to the similar event patterns recorded in historical data, identify current emergencies and generate corresponding response strategies, including real-time warning information, road closure instructions, and warning signal triggering plans; A response strategy generation unit, connected to the global evaluation unit, is used to receive the environmental prediction information, traffic management optimization suggestions, and emergency response strategies, and integrate these results into specific executable instructions; A result sending unit, connected to the response strategy generation unit, is used to transmit the environmental prediction information, traffic management optimization suggestions, and emergency response strategies to the Internet of Things communication module for use by terminal devices or traffic management centers; A model optimization unit, which is used to iteratively train the deep learning algorithm and the traffic flow evaluation algorithm based on the graph convolutional neural network and long short-term memory network based on the uploaded real-time data and historical data, as well as the actual execution feedback of the global evaluation result, to improve the prediction accuracy and response efficiency of the model output; A cloud storage unit, which is used to store historical environmental data, global evaluation results, and response strategies, and provide long-term data support for the global evaluation unit and the model optimization unit.

[0028] The global evaluation model includes: Environmental prediction sub - model: used to generate prediction information of the environment. The environmental prediction sub - model is constructed based on deep - learning algorithms, and its training and use include: Extract historical environmental data from the cloud storage unit as the training dataset; use the long short - term memory network as the basic model to capture the temporal changes of environmental data; optimize the model by minimizing the mean square error; Input real - time environmental data into the environmental prediction sub - model, and combine with historical data to generate environmental prediction information within a future time period; Traffic flow assessment sub - model: used to generate optimization suggestions for traffic management. The traffic flow assessment sub - model is constructed based on a combined algorithm of graph convolutional neural network and long short - term memory network, and its training and use include: Extract the traffic flow in the historical environmental data from the cloud storage unit and combine it with the road network topology as the training dataset; use the graph convolutional neural network to model the spatial features of the road network, and aggregate node features based on the following formula: , where A is the adjacency matrix of the road network, D is the degree matrix, is the feature matrix of the th layer, is the weight matrix, represents the activation function; Use the long short - term memory network to model the temporal features of traffic flow data, extract temporal features and combine with the output of the graph convolutional neural network; combine the results of the graph convolutional neural network and the long short - term memory network to generate a spatio - temporal joint representation of traffic flow, and optimize the model through the mean square error function; Input traffic flow and the current road network topology, combine with historical traffic data, generate an evaluation result of the current traffic flow state and a traffic flow prediction within a future time period; generate optimization suggestions for traffic management according to the prediction results; Emergency event recognition sub - model: used to generate response strategies for emergency events. The emergency event recognition sub - model is constructed based on a combined algorithm of convolutional neural network and temporal convolutional network in deep learning, and its training and use include: Extract historical emergency event data and corresponding environmental data from the cloud storage unit as the training dataset; use the convolutional neural network to extract the spatial features of environmental data and traffic flow data, and use the temporal convolutional network to model the temporal features of the data; combine the spatial features and temporal features to construct an emergency event recognition model, and optimize it through the cross - entropy loss function; Input real - time environmental data and traffic flow data to generate an emergency event detection result; generate corresponding response strategies according to the detection result.

[0029] The intelligent linkage module is respectively connected to the edge computing module and the Internet of Things communication module, and is used to receive real-time analysis results and global evaluation results, and execute corresponding linkage control operations, including dynamically adjusting the operating parameters of traffic management devices, triggering real-time warning devices or sending control instructions to the management center; The intelligent linkage module includes: A linkage data receiving unit, which is used to receive real-time analysis results from the edge computing module and global evaluation results from the cloud analysis module; A linkage strategy parsing unit, which is connected to the linkage data receiving unit, is used to parse the global evaluation results, and generate corresponding control instructions according to the parsing results; The parsing process includes: judging the meteorological change trend, air quality change and road condition prediction in the environmental prediction information to determine whether to trigger an environmental warning instruction; parsing the traffic signal cycle adjustment, diversion route planning and traffic guidance plan in the traffic management optimization suggestions to generate corresponding traffic signal device control instructions; parsing the road closure, real-time warning and warning signal triggering in the emergency response strategy to generate linkage control instructions; A device control unit, which is connected to the linkage strategy parsing unit, is used to execute the parsed control instructions; The device control includes: traffic signal device control, dynamically adjusting the traffic signal cycle, switching traffic signs or activating dynamic warning lights; warning device control, triggering warning broadcasts, electronic displays or other warning devices to provide real-time warning information to vehicles or personnel; A remote control module, which is used to send linkage control instructions to the traffic management center, and supports remote operation and manual intervention; A status feedback unit, which is used to monitor the device execution status in real time, and transmit the execution status and feedback information to the cloud analysis module and the device self-maintenance module for subsequent optimization and fault handling; A priority scheduling unit, which is connected to the linkage strategy parsing unit, is used to schedule multiple linkage control instructions according to the priority. When multiple linkage control tasks conflict, select and execute the linkage instructions according to the event type and urgency.

[0030] A power supply module, which is used to intelligently supply power to the highway environment monitoring terminal and support the low-power operation of each module; The power supply module includes: A solar power supply unit, which is used to capture light energy and convert it into electrical energy to provide sustainable energy support for the terminal; A storage battery unit, which is connected to the solar power supply unit, is used to store excess electrical energy and supply power to the terminal under insufficient light or no light conditions; A low-power management unit, which is used to dynamically adjust the power consumption status of each module, optimize power distribution by analyzing the operation status and energy usage requirements of the terminal; A backup power supply unit, which is connected to the energy storage battery unit and is used to provide emergency power supply when both solar power supply and energy storage battery are unavailable, ensuring the basic operation of the terminal; A power supply status monitoring unit, which is used to monitor the operation status of the solar power supply unit, the energy storage battery unit and the backup power supply unit in real time, including power, output power and operation efficiency, and transmit the monitoring results to the device self-maintenance module.

[0031] A device self-maintenance module, which is connected to the environmental monitoring module, the edge computing module, the Internet of Things communication module and the power supply module in the terminal, is used to monitor the operation status of each module, perform automatic detection and diagnosis of device faults, and transmit the fault diagnosis report to the intelligent linkage module.

[0032] The device self-maintenance module includes: A status monitoring unit, which is used to monitor the operation status of the environmental monitoring module, the edge computing module, the Internet of Things communication module, the power supply module and the intelligent linkage module in real time, including working status, data transmission status, power supply status and device connection status; A fault detection unit, which is connected to the status monitoring unit and is used to analyze the monitoring data and identify anomalies or faults in the module operation, including sensor failure, communication interruption, insufficient power and module hardware failure; A diagnosis unit, which is connected to the fault detection unit and is used to generate a fault diagnosis report based on the fault detection results, including fault cause, influence range and priority evaluation; A fault response unit, which is connected to the diagnosis unit and is used to automatically take corresponding response measures according to the fault diagnosis report, including restarting the module, switching the communication path, enabling the backup power supply or sending a fault alarm; A fault reporting unit, which is used to send the fault diagnosis report and response measures to the cloud analysis module or the traffic management center through the Internet of Things communication module; A log recording unit, which is used to record the device operation log and the fault diagnosis log, including time, location, fault type and processing result, and store the log in the cloud storage unit.

[0033] The above specific implementation manners further elaborate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above is only the specific implementation manners of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Highway environment monitoring terminal based on Internet of Things, characterized by: include: An environmental monitoring module is used to collect real-time environmental data along the highway through a sensor array, including meteorological parameters, air quality parameters, and road condition parameters. The environmental data is processed into structured information and transmitted to the edge computing module; The edge computing module is connected to the environmental monitoring module and processes the environmental data in real time based on the embedded real-time analysis model. The real-time analysis model includes an anomaly detection sub-model and a predictive analysis sub-model. By performing preliminary analysis on the environmental data, the real-time analysis results are generated, including data anomaly tags, trend prediction results and event response information, and the real-time analysis results are transmitted to the Internet of Things communication module and the intelligent linkage module. The IoT communication module is connected to the edge computing module and is used to upload environmental data and real-time analysis results to the cloud analysis module, and receive the global evaluation results returned by the cloud analysis module, so as to achieve interconnection between terminal devices and stable communication with the cloud platform; The cloud analysis module is connected to the edge computing module through the IoT communication module, and performs global analysis and processing on the environmental data and the real-time analysis results based on the global evaluation model in the cloud. The global evaluation model combines historical data with real-time data to generate a global evaluation result, including environmental prediction information, optimization suggestions for traffic management, and response strategies for emergencies, and transmits the global evaluation result back to the IoT communication module; The intelligent linkage module is connected to the edge computing module and the Internet of Things communication module respectively, and is used to receive real-time analysis results and global evaluation results, and perform corresponding linkage control operations, including dynamically adjusting the operating parameters of traffic management equipment, triggering real-time warning equipment, or sending control instructions to the management center.

2. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: The environmental monitoring module comprises: The sensor array includes: Meteorological sensors collect data related to meteorological conditions along the highway, the meteorological sensors include temperature sensors, humidity sensors and wind speed sensors, which are used to detect ambient temperature, humidity and wind speed respectively; Air quality sensors detect the concentration of pollutants in the air along the highway. The air quality sensors include PM2.5 sensors, PM10 sensors and carbon dioxide sensors, which are used to detect fine particulate matter, inhalable particulate matter and carbon dioxide concentration respectively; Road state sensing, detecting the state of the road surface, the road state sensing includes a camera, an infrared sensor and a humidity sensor, wherein the camera is used to collect image data of the road surface to analyze water accumulation, ice, obstacles and vehicle flow; the infrared sensor is used to detect the temperature distribution of the road surface to determine whether there is ice or temperature abnormality; the humidity sensor is used to detect the humidity level of the road surface to determine water accumulation or slippery conditions; The data processing unit is connected to the sensor array and is used to receive environmental data collected by various sensors, process the environmental data, remove abnormal data and noise data, and convert the processed data into structured information in a unified format and transmit it to the edge computing module.

3. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: The edge computing module includes: A data receiving unit, used for receiving environmental data from the environmental monitoring module; A data preprocessing unit, connected to the data receiving unit, for processing the received environmental data, including removing duplicate data and repairing lost data; A real-time analysis unit connected to the data preprocessing unit, for performing intelligent analysis on the processed data based on an embedded real-time analysis model, wherein the real-time analysis model includes an anomaly detection sub-model for detecting abnormal patterns or abnormal events in the data, including water accumulation, ice formation, or excessive air pollutants; a prediction analysis sub-model for predicting future trends of environmental parameters based on time series data, including road icing risk prediction and meteorological condition change trend prediction; The result generating unit is connected to the real-time analyzing unit, and is used for generating real-time analyzing results, and transmitting the generated real-time analyzing results to the Internet of Things communication module and the intelligent linkage module.

4. The highway environment monitoring terminal based on the Internet of Things according to claim 3 is characterized in that: The real-time analysis model includes: The anomaly detection sub-model uses an anomaly classification algorithm based on support vector machine to input environmental data is a feature vector, and a separation hyperplane in a high-dimensional space is constructed to determine whether the data is an abnormal point. The hyperplane is defined by the following formula: ,in is the weight vector, is the bias term, if , it is judged as normal data, otherwise it is judged as abnormal data; The anomaly detection sub-model is trained by extracting feature vectors from historical environmental data sets, with the classification accuracy of abnormal data and normal data as the optimization target; The trend prediction sub-model uses a time series prediction algorithm based on long short-term memory network to calculate the time series of environmental data {X t ,X t-1 ,…,X t-n } is input to predict environmental parameters at future time points; The trend prediction sub-model is trained by supervised learning, using historical time series data and its corresponding future target values ​​as training samples, and optimizing by minimizing mean square error as the loss function; The real-time analysis model is obtained through pre-training, and the training data comes from historical environmental monitoring data, including meteorological data, air quality data and road status data; the training process is completed through a distributed computing platform, including four stages: data preprocessing, feature extraction, model training and parameter optimization; the cross-validation method is used to verify the model performance, and classification accuracy, prediction error and model convergence are used as model evaluation criteria.

5. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: The Internet of Things communication module includes: A communication protocol unit, which is used to support wireless communication protocols, including LoRa, NB-IoT, 4G and 5G communication protocols, and automatically switch communication modes according to application scenarios, including low power consumption, high bandwidth and long-distance transmission; A data transmission unit, connected to the communication protocol unit, is used to upload the environmental data and real-time analysis results generated by the edge computing module to the cloud analysis module, and simultaneously receive the global evaluation results returned by the cloud analysis module; the data transmission unit encrypts the transmitted data; A data cache unit, connected to the data transmission unit, is used to temporarily store environmental data and real-time analysis results when communication is interrupted or the network is unstable, and automatically upload them after communication is restored; The communication status monitoring unit is used to monitor the operation status of the communication module in real time, including the network connection status, signal strength and data transmission rate, and send fault information to the equipment self-maintenance module when the communication is abnormal; The low-power unit is used to optimize the energy consumption of the communication module and dynamically adjust the communication power and working mode according to the data transmission task.

6. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: The cloud analysis module includes: A data receiving unit is used to receive environmental data and real-time analysis results uploaded by the IoT communication module, and to classify and store the received data, wherein the data is divided into real-time data and historical data according to the time dimension to support subsequent analysis; The global evaluation unit is connected to the data receiving unit, performs global analysis and processing on the received data based on the global evaluation model, and generates the following three types of results: Environmental prediction information: Based on deep learning algorithms, by extracting features from real-time environmental data and combining the temporal changes of historical environmental data, future environmental conditions can be predicted, including weather trends, air quality distribution, and future changes in road conditions. Traffic management optimization suggestions: Based on graph convolutional neural networks and long short-term memory networks, by jointly modeling real-time traffic flow data and historical traffic data, the dynamic change characteristics of traffic flow are analyzed to generate traffic management optimization suggestions, including traffic light cycle adjustment, diversion route planning and traffic diversion plans; Emergency response strategy: Identify current emergencies through real-time environmental data and refer to similar event patterns recorded in historical data, and generate corresponding response strategies, including real-time warning information, road closure instructions and warning signal triggering plans; A response strategy generation unit, connected to the global evaluation unit, is used to receive environmental prediction information, traffic management optimization suggestions and emergency response strategies, and integrate these results into specific executable instructions; A result sending unit, connected to the response strategy generating unit, is used to transmit the environmental prediction information, traffic management optimization suggestions and emergency response strategies to the Internet of Things communication module for use by the terminal device or the traffic management center; Model optimization unit, which is used to iteratively train deep learning algorithms, graph convolutional neural networks, and long short-term memory network-based traffic flow assessment algorithms based on uploaded real-time and historical data, as well as actual execution feedback of global assessment results, to improve the prediction accuracy and response efficiency of model output; The cloud storage unit is used to store historical environmental data, global assessment results and response strategies, and provide long-term data support for the global assessment unit and model optimization unit.

7. The highway environment monitoring terminal based on the Internet of Things according to claim 6 is characterized in that: The global evaluation model includes: Environmental prediction sub-model: used to generate prediction information of the environment. The environmental prediction sub-model is built based on a deep learning algorithm. Its training and use include: Extract historical environmental data from cloud storage units as training data sets; use long short-term memory networks as the basic model to capture the temporal changes of environmental data; optimize the model by minimizing the mean square error; Input the real-time environmental data into the environmental prediction sub-model and combine it with historical data to generate environmental prediction information for future time periods; The traffic flow assessment sub-model is used to generate optimization suggestions for traffic management. The traffic flow assessment sub-model is constructed based on a combined algorithm of a graph convolutional neural network and a long short-term memory network. Its training and use include: Extract the traffic volume in the historical environmental data from the cloud storage unit and combine it with the road network topology as a training data set; use the graph convolutional neural network to model the spatial characteristics of the road network, use the long short-term memory network to model the temporal characteristics of the traffic flow data, extract the time series features and combine them with the output of the graph convolutional neural network; combine the results of the graph convolutional neural network and the long short-term memory network to generate a spatial-temporal joint representation of the traffic flow, and optimize the model through the mean square error function; Input the traffic volume and current road network topology, combine with historical traffic data, generate the evaluation results of the current traffic flow status and the traffic flow forecast in the future time period; generate optimization suggestions for traffic management based on the forecast results; The emergency event recognition sub-model is used to generate a response strategy for an emergency event. The emergency event recognition sub-model is constructed based on a combination algorithm of a convolutional neural network and a temporal convolutional network in deep learning. Its training and use include: Extract historical emergency data and corresponding environmental data from cloud storage units as training data sets; use convolutional neural networks to extract spatial features of environmental data and traffic flow data, and use temporal convolutional networks to model the temporal features of data; combine spatial features and temporal features to build an emergency recognition model, and optimize it through the cross entropy loss function; Input real-time environmental data and traffic flow data to generate emergency detection results; generate corresponding response strategies based on the detection results.

8. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: The intelligent linkage module comprises: A linkage data receiving unit, used to receive real-time analysis results from the edge computing module and global evaluation results from the cloud analysis module; The linkage strategy parsing unit is connected to the linkage data receiving unit and is used to parse the global evaluation result and generate corresponding control instructions according to the parsing result; The analysis process includes: judging the meteorological change trend, air quality change and road status forecast in the environmental forecast information to decide whether to trigger the environmental warning instruction; analyzing the traffic light cycle adjustment, diversion route planning and traffic diversion plan in the traffic management optimization suggestions to generate corresponding traffic signal equipment control instructions; analyzing the road closure, real-time warning and warning signal triggering in the emergency response strategy to generate linkage control instructions; A device control unit, connected to the linkage strategy parsing unit, for executing the parsed control instructions; The equipment control includes: traffic signal equipment control, dynamically adjusting the traffic signal light cycle, switching traffic signs or enabling dynamic warning lights; early warning equipment control, triggering early warning broadcasts, electronic display screens or other warning equipment to provide real-time early warning information to vehicles or personnel; Remote control module, used to send linkage control instructions to the traffic management center, supporting remote operation and manual intervention; The status feedback unit is used to monitor the execution status of the equipment in real time and transmit the execution status and feedback information to the cloud analysis module and the equipment self-maintenance module for subsequent optimization and fault handling; The priority scheduling unit is connected to the linkage strategy analysis unit and is used to schedule multiple linkage control instructions according to priority. When multiple linkage control tasks conflict, the linkage instruction is selected for execution according to the event type and urgency.

9. The highway environment monitoring terminal based on the Internet of Things according to claim 1 is characterized in that: Also includes: The power module is used to intelligently power the highway environment monitoring terminal and support low-power operation of each module; The power module comprises: A solar power supply unit, which is used to capture light energy and convert it into electrical energy, providing sustainable energy support for the terminal; An energy storage battery unit, connected to the solar power supply unit, is used to store excess electrical energy and to supply power to the terminal under insufficient or no light conditions; Low power management unit, used to dynamically adjust the power consumption status of each module, optimize power distribution by analyzing the terminal's operation status and energy usage requirements; A backup power supply unit, connected to the energy storage battery unit, is used to provide emergency power supply when both the solar power supply and the energy storage battery are unavailable, thereby ensuring the basic operation of the terminal; The power status monitoring unit is used to monitor the operating status of the solar power supply unit, energy storage battery unit and backup power supply unit in real time, including power, output power and operating efficiency, and transmit the monitoring results to the equipment self-maintenance module.

10. The highway environment monitoring terminal based on the Internet of Things according to claim 1, characterized in that: Also includes: The equipment self-maintenance module is connected to the environment monitoring module, edge computing module, IoT communication module and power module in the terminal to monitor the operating status of each module, perform automatic detection and diagnosis of equipment faults, and transmit fault diagnosis reports to the intelligent linkage module; The equipment self-maintenance module comprises: The status monitoring unit is used to monitor the operating status of the environmental monitoring module, edge computing module, IoT communication module, power module and intelligent linkage module in real time, including working status, data transmission status, power status and device connection status; The fault detection unit is connected to the status monitoring unit and is used to analyze the monitoring data and identify abnormalities or faults in the operation of the module, including sensor failure, communication interruption, insufficient power and module hardware failure; A diagnosis unit connected to the fault detection unit, for generating a fault diagnosis report based on the fault detection result, including the cause of the fault, the scope of impact and the priority assessment; A fault response unit, connected to the diagnosis unit, for automatically taking corresponding response measures according to the fault diagnosis report, including restarting the module, switching the communication path, enabling the backup power supply, or sending a fault alarm; A fault reporting unit, used to send fault diagnosis reports and response measures to a cloud analysis module or a traffic management center through an IoT communication module; The log recording unit is used to record the equipment operation log and fault diagnosis log, including time, location, fault type and processing results, and store the log in the cloud storage unit.

Images