Dynamic Traffic Environment Monitoring Method and Device

Abstract

The embodiments of the present application provide a method and device for dynamic traffic environment monitoring. By constructing an environmental monitoring and collection network for power-assisted bicycles, mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide is achieved. The cloud computing platform identifies outliers based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. A deep reinforcement learning algorithm is used to realize the intelligent scheduling of monitoring resources, and an optimized inspection plan is generated according to the pollution risk level, vehicle location, and battery power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to a method and device for monitoring dynamic traffic environment. Background Art

[0002] The existing traffic environment monitoring methods have obvious deficiencies. Traditional systems mainly rely on fixed monitoring stations, with low distribution density and difficulty in covering all areas, and cannot achieve dynamic and refined monitoring of urban air quality.

[0003] In addition, there are bottlenecks in data analysis and prediction in the existing technology. Most systems lack in-depth exploration of the spatio-temporal distribution characteristics of pollutants and cannot accurately predict the pollutant diffusion trend, resulting in insufficient pertinence of early warning and control measures.

[0004] There are technical shortcomings in the monitoring resource scheduling of existing systems. Lack of an intelligent task allocation mechanism, it is difficult to achieve efficient utilization of monitoring resources according to the pollution risk level. Solving these problems is of great significance for improving the accuracy and effectiveness of traffic environment monitoring. Summary of the Invention

[0005] In view of the problems in the existing technology, this application provides a method and device for monitoring dynamic traffic environment, which can effectively solve the deficiencies of traditional technology in monitoring coverage, data analysis and resource scheduling, and significantly improve the accuracy and practical value of traffic environment monitoring.

[0006] To solve at least one of the above problems, this application provides the following technical solutions:

[0007] In a first aspect, this application provides a method for monitoring dynamic traffic environment, including:

[0008] Configuring an assisted bicycle collection network with environmental monitoring collection devices, the environmental monitoring collection devices collect particulate matter concentration, nitrogen oxide content and carbon monoxide concentration, record the sampling position coordinates and sampling timestamps, and the environmental monitoring collection devices organize the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates and the sampling timestamps into monitoring data packets and upload them to the cloud computing platform;

[0009] The cloud computing platform performs data preprocessing on the received monitoring data packets, identifies outliers for the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix from the effective monitoring data according to the sampling location coordinates and the sampling timestamps, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform combines historical monitoring data to train a pollutant migration model, inputs the output features of the pollutant concentration distribution model into the pollutant migration model, and generates a pollutant diffusion trend prediction result;

[0010] The cloud computing platform divides pollution risk level regions according to the pollutant diffusion trend prediction result, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, uses the pollution risk level regions, the real-time location information and remaining power of the power-assisted bicycles as input parameters to generate a monitoring task allocation plan. The cloud computing platform sends the monitoring task allocation plan including the inspection route to the power-assisted bicycles, calculates the pollution risk index for each road section in the future period according to the pollutant diffusion trend prediction result, and generates a warning message including traffic control strategies.

[0011] Further, it also includes: dividing urban roads into regions based on road traffic flow distribution and wireless communication signal strength. The cloud computing platform calculates the monitoring point density threshold for each region according to the traffic flow density, road type, and passing time period, determines the number of power-assisted bicycles required in each region based on the monitoring point density threshold, generates acquisition task configuration parameters including the sampling period and monitoring duration, and allocates the acquisition task configuration parameters to the power-assisted bicycles;

[0012] The particulate matter sensor in the environmental monitoring and acquisition device detects fine particulate matter in the air based on the laser scattering principle. The nitrogen oxide sensor obtains the nitrogen oxide content by generating an electrical signal through the reaction of electrode materials. The carbon monoxide sensor measures the carbon monoxide concentration according to the change in the resistance of the metal oxide semiconductor material. The environmental monitoring and acquisition device obtains the sampling location coordinates based on the Beidou navigation and positioning module, and associates the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration with the sampling location coordinates to generate monitoring data.

[0013] Further, it further includes: the environmental monitoring and acquisition device sets a cache data pool to store the acquired monitoring data. The particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration in the cache data pool respectively establish a mapping relationship with the sampling position coordinates and the sampling timestamp. The environmental monitoring and acquisition device reads the monitoring data from the cache data pool according to a preset data cache threshold, performs standardization processing and coding compression on the read monitoring data, and generates the monitoring data packet;

[0014] The environmental monitoring and acquisition device performs format encapsulation on the monitoring data packet based on a preset data transmission protocol, selects a transmission channel according to the wireless network signal quality. The environmental monitoring and acquisition device uploads the encapsulated monitoring data packet to the cloud computing platform through the transmission channel. The cloud computing platform parses and performs integrity verification on the received monitoring data packet, and stores the monitoring data that passes the verification in the database.

[0015] Further, it further includes: the cloud computing platform respectively performs normalization processing on the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration in the monitoring data packet, calculates the data similarity of adjacent sampling points by using a time series clustering algorithm, marks the data points with a similarity deviation greater than a preset threshold as outliers. The cloud computing platform corrects the marked outliers based on a linear interpolation method, and uses the corrected monitoring data as the effective monitoring data;

[0016] The cloud computing platform arranges the effective monitoring data in time series according to the sampling timestamp, maps the effective monitoring data within the same time period to a two-dimensional spatial grid based on the sampling position coordinates, uses the Kriging interpolation method to supplement the data of the missing measurement points in the two-dimensional spatial grid, and constructs the supplemented grid data into the spatio-temporal data matrix. The row vector of the spatio-temporal data matrix represents the sampling time, and the column vector represents the spatial position.

[0017] Further, it further includes: the cloud computing platform inputs the spatio-temporal data matrix into a multi-channel convolutional neural network. The multi-channel convolutional neural network respectively extracts features in the time dimension and the spatial dimension by using a two-channel structure. The features in the time dimension extract the temporal change pattern of the pollutant concentration through a one-dimensional convolutional layer, and the features in the spatial dimension extract the spatial distribution law of the pollutant concentration through a two-dimensional convolutional layer. The features in the time dimension and the features in the spatial dimension are fused to obtain the spatio-temporal features of the pollutant, and a pollutant concentration distribution model is constructed based on the spatio-temporal features of the pollutant;

[0018] The cloud computing platform reads historical monitoring data from the database, clusters the historical monitoring data according to pollutant types and concentration change trends, and uses a long short-term memory network to establish the pollutant migration model. The pollutant migration model takes the output features of the pollutant concentration distribution model as input, combines meteorological data to predict the diffusion trend of pollutants at different spatio-temporal scales, and generates a spatio-temporal evolution sequence of the pollutant concentration field based on the diffusion trend.

[0019] Further, it also includes: The cloud computing platform calculates the pollutant concentration gradient and change rate of each region based on the pollutant diffusion trend prediction result, inputs the pollutant concentration gradient and the change rate into a fuzzy neural network. The fuzzy neural network divides the risk level of the monitoring area according to a preset pollutant concentration threshold, generates a pollution risk distribution map including high, medium, and low three-level risk levels, and the cloud computing platform performs grid processing on the pollution risk distribution map to obtain the pollution risk level area.

[0020] The cloud computing platform constructs the monitoring resource scheduling model using a deep reinforcement learning algorithm. The monitoring resource scheduling model takes the pollution risk level area as the state space, takes the real-time position information and remaining power of the power-assisted bicycle as constraint conditions, calculates the optimal monitoring path based on the value function iteration method, and the monitoring resource scheduling model assigns inspection routes and sampling frequencies to each power-assisted bicycle according to the optimal monitoring path, generating the monitoring task allocation plan including the sampling area and sampling duration.

[0021] Further, it also includes: The cloud computing platform encodes the monitoring task allocation plan according to the data transmission protocol, and issues the encoded monitoring task allocation plan to each power-assisted bicycle based on the wireless communication network. After receiving the monitoring task allocation plan, the power-assisted bicycle decodes it, plans the sampling trajectory according to the inspection route, and the cloud computing platform real-time tracks the task execution status of each power-assisted bicycle. When it detects that the power-assisted bicycle deviates from the inspection route, it recalculates the optimal monitoring path.

[0022] The cloud computing platform performs grid division on the road network in the monitoring area based on the pollutant diffusion trend prediction result, calculates the pollutant concentration change trend of each grid unit in the future time period, compares the pollutant concentration change trend with a preset pollutant concentration threshold to obtain the pollution risk index, and the cloud computing platform determines the traffic flow regulation area according to the spatial distribution of the pollution risk index, and generates the warning information including the vehicle flow guidance plan based on the correlation model between traffic flow and pollutant concentration.

[0023] In a second aspect, the present application provides a dynamic traffic environment monitoring device, including:

[0024] An environmental acquisition module is used to configure an assisted bicycle acquisition network with environmental monitoring acquisition devices. The environmental monitoring acquisition devices collect particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, record the sampling position coordinates and sampling timestamps, and organize the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into monitoring data packets and upload them to a cloud computing platform;

[0025] A pollution prediction module is used for the cloud computing platform to perform data preprocessing on the received monitoring data packets, identify outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on a time series clustering algorithm, obtain effective monitoring data, the cloud computing platform constructs a spatio-temporal data matrix according to the sampling position coordinates and the sampling timestamps for the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, establishes a pollutant concentration distribution model, the cloud computing platform trains a pollutant migration model in combination with historical monitoring data, and inputs the output features of the pollutant concentration distribution model into the pollutant migration model to generate a pollutant diffusion trend prediction result;

[0026] A dynamic monitoring module is used for the cloud computing platform to divide pollution risk level regions according to the pollutant diffusion trend prediction result, construct a monitoring resource scheduling model based on a deep reinforcement learning algorithm, use the pollution risk level regions, the real-time position information and remaining power of the assisted bicycle as input parameters to generate a monitoring task allocation plan, the cloud computing platform sends the monitoring task allocation plan including the inspection route to the assisted bicycle, and calculates the pollution risk index of each road section in the future period according to the pollutant diffusion trend prediction result to generate a warning message including traffic control strategies.

[0027] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the dynamic traffic environment monitoring method are implemented.

[0028] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the dynamic traffic environment monitoring method are implemented.

[0029] In a fifth aspect, the present application provides a computer program product, including a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of the dynamic traffic environment monitoring method are implemented.

[0030] As can be seen from the above technical solutions, the present application provides a method and device for monitoring a dynamic traffic environment. By constructing a monitoring and collection network for the assisted bicycle environment, mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide is achieved. The cloud computing platform identifies outliers based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. The intelligent scheduling of monitoring resources is realized by using a deep reinforcement learning algorithm, and an optimized inspection plan is generated according to the pollution risk level, vehicle location, and battery power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0032] Figure 1 It is a schematic flowchart of the method for monitoring a dynamic traffic environment in an embodiment of the present application;

[0033] Figure 2 It is a structural diagram of the device for monitoring a dynamic traffic environment in an embodiment of the present application;

[0034] Figure 3 It is a schematic structural diagram of an electronic device in an embodiment of the present application.

[0035] Reference numerals:

[0036] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present application belong to the scope of protection of the present application.

[0038] In the technical solution of this application, the acquisition, storage, use, processing, etc. of data all comply with the relevant provisions of national laws and regulations.

[0039] Considering the problems existing in the prior art, this application provides a dynamic traffic environment monitoring method and device. By constructing a monitoring and collection network for the assisted bicycle environment, mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide is realized. The cloud computing platform identifies outliers based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. The intelligent scheduling of monitoring resources is realized by using a deep reinforcement learning algorithm, and an optimized inspection plan is generated according to the pollution risk level, vehicle position, and battery power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0040] In order to effectively solve the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improve the accuracy and practical value of traffic environment monitoring, this application provides an embodiment of a dynamic traffic environment monitoring method. Refer to Figure 1 , the dynamic traffic environment monitoring method specifically includes the following contents:

[0041] Step S101: Configure an assisted bicycle collection network with an environmental monitoring and collection device. The environmental monitoring and collection device collects particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling position coordinates and sampling timestamps, and the environmental monitoring and collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into a monitoring data packet and uploads it to the cloud computing platform;

[0042] Optionally, this embodiment designs a mobile environmental monitoring network based on an assisted bicycle. The assisted bicycle adopts a modular design, and on the premise of not affecting the riding experience, the environmental monitoring and collection device is integrated inside the handlebar riser. The collection device adopts a low-power design and is continuously powered through a bicycle power generation system and a lithium battery pack to ensure the long-term stable operation of the monitoring equipment.

[0043] This embodiment innovatively realizes a multi-parameter collaborative monitoring mechanism. The particulate matter sensor is based on the principle of laser scattering, uses an infrared laser to emit a beam, and calculates the particulate matter concentration by detecting the scattering intensity of light by suspended particulate matter in the air. The sensor adopts an air flow channel design to actively guide air flow and improve the sampling efficiency. At the same time, the accuracy of particulate matter particle size identification is improved through multi-angle scattered light detection.

[0044] This embodiment optimizes the gas pollutant detection method. The nitrogen oxide sensor adopts the electrochemical principle, and generates an electrical signal through the redox reaction between a specific electrode material and the target gas. The sensor is built-in with a temperature compensation circuit to eliminate the influence of ambient temperature on the measurement result. The carbon monoxide sensor uses metal oxide semiconductor materials and measures the concentration by detecting the resistance change caused by gas adsorption. Both sensors adopt anti-interference design to reduce the cross-influence of other gases.

[0045] This embodiment realizes the acquisition of accurate position information. The acquisition device integrates Beidou and GPS dual-mode positioning modules, supporting multi-constellation collaborative positioning. The satellite positioning data and inertial sensor data are fused through the Kalman filtering algorithm to improve the positioning accuracy. The network time protocol is used for time synchronization to ensure the accuracy of the sampling timestamp. The position coordinates adopt the WGS84 coordinate system, which is convenient for subsequent data processing and analysis.

[0046] This embodiment designs a data preprocessing and packetizing mechanism. The environmental monitoring acquisition device standardizes the original sensor data to eliminate the range difference. The data packetizing adopts a hierarchical structure: the basic information layer includes the device ID and timestamp; the monitoring data layer includes the concentration values of various pollutants; the position information layer includes the longitude and latitude coordinates. The data compression algorithm is used to reduce the transmission load while ensuring the data accuracy.

[0047] This embodiment constructs a reliable data transmission mechanism. The acquisition device supports 4G / 5G mobile communication and Wi-Fi dual-mode transmission, and automatically selects the optimal transmission channel according to the network conditions. The MQTT protocol is used for data transmission to achieve reliable message push. At the same time, a local caching mechanism is implemented to temporarily store data when the network is unavailable and automatically retransmit it when the network resumes.

[0048] This embodiment optimizes the deployment strategy of the acquisition network. By analyzing the urban road traffic flow distribution and the wireless network coverage, the monitoring area is divided into multiple sub-areas. Each sub-area is configured with a corresponding number of assisted bicycles according to the area size and pollutant distribution characteristics. The sampling density is dynamically adjusted according to the regional pollution risk level to ensure the reasonable allocation of monitoring resources.

[0049] This embodiment realizes adaptive sampling control. The acquisition device dynamically adjusts the sampling frequency according to the riding speed, increasing the sampling frequency in the low-speed or stopped state and appropriately reducing the frequency during high-speed riding to ensure the data quality. At the same time, the key monitoring areas and periods are determined by combining historical data analysis to optimize the sampling strategy.

[0050] This embodiment designs a complete data quality assurance system. The sensors are calibrated regularly to ensure measurement accuracy. Through a multiple anomaly detection mechanism, abnormal data is identified and marked. The self-diagnosis function of sensor faults is realized, and maintenance personnel are automatically notified when sensor anomalies are detected. At the same time, real-time validity verification is performed on the collected data to ensure the reliability of the uploaded data.

[0051] Through the above technological innovations, this embodiment solves the problems of limited coverage and low spatio-temporal resolution of traditional fixed environmental monitoring stations. In practical applications, this solution makes full use of the urban shared bicycle system to build a large-scale and low-cost mobile environmental monitoring network. Through multi-sensor collaborative monitoring and intelligent data processing, refined monitoring of urban pollutant distribution is achieved, providing reliable data support for urban environmental management and pollution prevention and control.

[0052] Step S102: The cloud computing platform performs data preprocessing on the received monitoring data packets, identifies outliers of the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration based on the time series clustering algorithm, obtains effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix according to the sampling position coordinates and the sampling timestamps of the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform trains a pollutant migration model in combination with historical monitoring data, and inputs the output features of the pollutant concentration distribution model into the pollutant migration model to generate a pollutant diffusion trend prediction result;

[0053] Optionally, this embodiment innovatively realizes the preprocessing process of monitoring data. The cloud computing platform first unpacks and extracts parameters from the received monitoring data packets, and eliminates the range differences between different sensors through data standardization. The time series data preprocessing adopts a sliding window mechanism, and the window size is dynamically adjusted according to the data sampling frequency to ensure capturing the time characteristics of the pollutant concentration.

[0054] This embodiment designs an outlier recognition method based on time series clustering. The improved DBSCAN algorithm is used to perform clustering analysis on time series data, and the clustering features include pollutant concentration values and change rates. The algorithm identifies outliers through the density reachability criterion, and the calculation formula is:

[0055] D(p) = Σ(dist(p,q)) / N, q∈Nε(p)

[0056] where D(p) is the local density of data point p, dist(p,q) is the distance between point p and point q in the neighborhood, Nε(p) is the set of data points within the ε neighborhood of point p, and N is the number of points in the neighborhood. When D(p) is significantly greater than the average density of the neighborhood, p is marked as an outlier.

[0057] This embodiment optimizes the construction method of the spatio-temporal data matrix. Based on the sampling position coordinates, the monitoring area is divided into regular grids, and the grid size is determined according to the sampling density. The Kriging interpolation algorithm is used to supplement the missing measurement points within the grids, and the spatial autocorrelation is considered to improve the interpolation accuracy. In the time dimension, equidistant sampling is adopted to ensure the temporal continuity of the data. In the constructed matrix, the rows represent the time series, the columns represent the spatial positions, and the matrix elements are the pollutant concentration values at the corresponding spatio-temporal points.

[0058] This embodiment innovatively designs a multi-channel convolutional neural network structure. The network adopts a two-channel architecture. The time channel uses a one-dimensional convolutional layer to extract the temporal variation patterns of pollutant concentrations, and the spatial channel uses a two-dimensional convolutional layer to capture the spatial distribution characteristics of pollutants. The features of the two channels are fused through an attention mechanism to highlight important spatio-temporal features. The network training adopts a transfer learning strategy to accelerate convergence using a pre-trained model.

[0059] This embodiment optimizes the training process of the pollutant concentration distribution model. The model input is the spatio-temporal data matrix, and hierarchical features are extracted through multi-layer convolutional operations. The intermediate layer adopts a residual connection structure to alleviate the problem of gradient disappearance. The model output includes the spatial distribution map and the temporal variation trend of pollutant concentrations. During the training process, a dynamic learning rate adjustment strategy is adopted to improve the model convergence efficiency.

[0060] This embodiment implements an innovative pollutant migration model. A time series prediction model is constructed based on the Long Short-Term Memory network (LSTM), and the input features include pollutant concentration distribution, meteorological parameters, and geographical information. The model learns the long-term dependencies of pollutant diffusion through a gating mechanism and captures the migration laws at different time scales. At the same time, an attention mechanism is introduced to dynamically adjust the feature weights according to the correlation of historical data.

[0061] This embodiment designs a model fusion prediction mechanism. The output features of the pollutant concentration distribution model are input into the migration model to achieve multi-scale prediction. Short-term prediction mainly relies on real-time monitoring data, while medium- and long-term prediction combines historical data and meteorological conditions. The prediction results include the spatio-temporal evolution sequence of the pollutant concentration field, which can be used to evaluate the pollutant diffusion trend.

[0062] This embodiment constructs a complete model evaluation system. The cross-validation method is used to evaluate the model performance, and the validation metrics include prediction accuracy, spatio-temporal correlation, etc. The adaptability of the model in different scenarios is verified through comparative experiments. At the same time, a model update mechanism is established to regularly retrain the model using newly collected data to maintain the prediction performance.

[0063] This embodiment realizes the accurate analysis and prediction of urban pollutant distribution through a data-driven approach. In practical applications, this solution can accurately identify abnormal pollution events, predict the diffusion trend of pollutants, and provide a scientific basis for environmental management decisions. Especially in the fields of air quality early warning and pollution prevention and control, the system significantly improves the accuracy and timeliness of prediction through multi-source data fusion and deep learning models.

[0064] The technical solution designed in this embodiment solves problems such as poor data quality and low prediction accuracy in traditional environmental monitoring. Through intelligent data processing and deep learning models, high-precision spatial interpolation and time-series prediction of pollutant concentrations are achieved, providing strong support for urban environmental quality assessment and early warning. The innovation of this solution is reflected in multiple aspects such as data processing, model construction, and prediction methods, and has important practical application value.

[0065] Step S103: The cloud computing platform divides the pollution risk level areas according to the prediction result of the pollutant diffusion trend, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, takes the pollution risk level areas, the real-time position information and remaining power of the assisted bicycle as input parameters, generates a monitoring task allocation plan, the cloud computing platform sends the monitoring task allocation plan including the inspection route to the assisted bicycle, calculates the pollution risk index of each road section in the future time period according to the prediction result of the pollutant diffusion trend, and generates an early warning information including traffic control strategies.

[0066] Optionally, this embodiment realizes intelligent risk zoning management based on the prediction result of the pollutant diffusion trend. The cloud computing platform first calculates the spatial gradient and time change rate of the pollutant concentration, and constructs a risk assessment model through a fuzzy neural network. The risk index R is calculated by the following formula:

[0067] R = w1G + w2V + w3 C

[0068] Where G is the pollutant concentration gradient, V is the concentration change rate, C is the current concentration value, and w1, w2, and w3 are weight coefficients, which are dynamically adjusted according to the harm degree of different pollutants.

[0069] This embodiment innovatively designs a deep reinforcement learning-based monitoring resource scheduling model. The model adopts a dual situation learning framework, maps the pollution risk level areas to the state space, and takes the position and power of the assisted bicycle as constraint conditions. The action space includes inspection route selection and sampling frequency adjustment. The design of the reward function considers multiple factors such as monitoring coverage, sampling efficiency, and resource consumption, and evaluates the long-term benefits of different scheduling strategies through a value network.

[0070] This embodiment optimizes the task allocation algorithm. Based on the A algorithm, it calculates the optimal path for the assisted bicycle to each monitoring point. The path weight comprehensively considers distance, road conditions, and risk levels. For high-risk areas, the system preferentially allocates multiple vehicles to cooperate in monitoring to improve data reliability. At the same time, a task dynamic adjustment mechanism is implemented. When a new pollution hotspot is detected, the system automatically re-plans the inspection route.

[0071] This embodiment implements an adaptive sampling strategy. It increases the sampling frequency in high-risk areas and optimizes the sampling point distribution through a dynamic programming algorithm. The calculation formula for the sampling frequency F is:

[0072] F = Fbase (αR + βD + γ E)

[0073] where Fbase is the reference sampling frequency, R is the risk index, D is the sampling point density, E is the remaining battery power, and α, β, and γ are adjustment coefficients. This adaptive mechanism ensures the efficient utilization of monitoring resources.

[0074] This embodiment designs an intelligent task distribution mechanism. The monitoring task allocation scheme adopts a hierarchical structure, including a basic configuration layer and a task execution layer. The basic configuration layer defines device parameters and sampling requirements, and the task execution layer includes specific inspection routes and sampling strategies. The system sends tasks to the assisted bicycle through a secure communication channel and real-time tracks the task execution status.

[0075] This embodiment constructs a real-time task monitoring system. The cloud platform uses GPS positioning to real-time track the position of the assisted bicycle and triggers path re-planning when it detects that the vehicle deviates from the predetermined route. At the same time, it monitors the battery power status of the device and automatically adjusts the task allocation when the power is insufficient to ensure the continuity of the monitoring work. The system also establishes a task completion degree evaluation mechanism to optimize subsequent scheduling strategies.

[0076] This embodiment innovatively implements a pollution risk index calculation method. Based on the pollutant diffusion prediction results, the system divides the monitoring area into grids and calculates the risk index of each grid at different time scales. By combining traffic flow data analysis of the pollutant accumulation effect, it establishes an association model between traffic activities and air quality. This multi-dimensional risk assessment provides a scientific basis for traffic control decisions.

[0077] This embodiment optimizes the warning information generation strategy. The system determines the traffic control area according to the spatial distribution of the risk index and evaluates the effects of different control schemes through traffic flow simulation. The warning information includes content such as area control suggestions, vehicle flow guidance plans, and detour suggestions. The information release adopts a multi-level trigger mechanism to ensure the timeliness and pertinence of the warning.

[0078] In this embodiment, through the intelligent scheduling and risk warning of monitoring resources, the problems of many coverage blind spots and response lags in traditional environmental monitoring are effectively solved. In practical applications, this solution can flexibly adjust the monitoring strategy according to the pollutant diffusion trend to achieve the optimal allocation of monitoring resources. Especially in the response to sudden pollution incidents, the system significantly improves the efficiency of environmental monitoring and the accuracy of early warning through real-time task adjustment and multi-level early warning.

[0079] The technical innovation points of this embodiment are mainly reflected in aspects such as the deep reinforcement learning scheduling model, adaptive sampling strategy, and multi-dimensional risk assessment. Through artificial intelligence algorithms, the intelligent scheduling of monitoring resources is realized, greatly improving the accuracy and timeliness of environmental monitoring. At the same time, the traffic control strategy based on pollution diffusion prediction provides strong support for urban refined management. This solution has important practical significance in improving urban air quality and protecting public health.

[0080] As can be seen from the above description, the dynamic traffic environment monitoring method provided by the embodiments of this application can realize the mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide by constructing an environmental monitoring collection network for assisted bicycles. The cloud computing platform identifies outliers based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with the pollutant migration model. The intelligent scheduling of monitoring resources is realized by using the deep reinforcement learning algorithm, and an optimized inspection plan is generated according to the pollution risk level, vehicle location, and battery power. This method effectively solves the deficiencies of traditional technologies in monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0081] In an embodiment of the dynamic traffic environment monitoring method of this application, the following content may also be specifically included:

[0082] Step S201: Divide urban roads based on the road traffic flow distribution and wireless communication signal strength. The cloud computing platform calculates the monitoring point density threshold for each area according to the traffic flow density, road type, and passing period, determines the number of assisted bicycles required in each area based on the monitoring point density threshold, generates collection task configuration parameters including the sampling period and monitoring duration, and allocates the collection task configuration parameters to the assisted bicycles;

[0083] Step S202: The particulate matter sensor in the environmental monitoring and collection device detects fine particulate matter in the air based on the principle of laser scattering. The nitrogen oxide sensor obtains the nitrogen oxide content by generating an electrical signal through the reaction of the electrode material. The carbon monoxide sensor measures the carbon monoxide concentration according to the change in the resistance of the metal oxide semiconductor material. The environmental monitoring and collection device obtains the sampling position coordinates based on the Beidou navigation and positioning module, and associates the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration with the sampling position coordinates to generate monitoring data.

[0084] Optionally, in this embodiment, an intelligent regional division strategy is implemented based on multi-dimensional data. The cloud computing platform first collects road traffic flow data, including information such as the spatio-temporal distribution of vehicle flow and the characteristics of peak hours. At the same time, it obtains the wireless communication network coverage map and analyzes the communication quality through the signal strength heat map. The regional division uses an improved K-means clustering algorithm to organize road segments with similar characteristics into monitoring units, ensuring that the division results meet the monitoring requirements and communication constraints.

[0085] In this embodiment, a monitoring point density calculation method is innovatively designed. The monitoring point density threshold D for each region is calculated by the following formula:

[0086] D = Dbase (w1F + w2R + w3 T)

[0087] where Dbase is the reference density value, F is the normalized vehicle flow density, R is the road grade coefficient, T is the time period weight, and w1, w2, and w3 are adjustment factors. This adaptive density calculation method can dynamically adjust the monitoring intensity according to the regional characteristics.

[0088] In this embodiment, the configuration strategy of the number of assisted bicycles is optimized. Based on the monitoring point density threshold, the system calculates the minimum coverage number required for each region. Considering the device availability rate and task overlap degree, the actual configuration number is adjusted by the redundancy coefficient. At the same time, a device dynamic allocation mechanism is established, which can flexibly adjust the number of vehicles between regions according to the monitoring requirements.

[0089] In this embodiment, intelligent task configuration generation is realized. The sampling period is determined according to the pollutant characteristics and the regional risk level, and a shorter sampling interval is adopted for high-risk regions. The monitoring duration is determined based on historical data analysis, considering the daily variation law of the pollutant concentration. The configuration parameters adopt a hierarchical structure, including basic configuration and scenario configuration, which can be flexibly adjusted according to the actual situation.

[0090] This embodiment designs an innovative particulate matter detection mechanism. The laser scattering sensor adopts the 90-degree scattered light detection principle, and improves the particle size recognition accuracy through the analysis of multi-angle scattered light signals. The sensor is built-in with a temperature and humidity compensation circuit to eliminate the influence of environmental factors. The sampling air flow channel is optimized to improve the capture efficiency of fine particulate matter.

[0091] This embodiment optimizes the gas pollutant detection method. The nitrogen oxide sensor adopts a dual-electrode structure, and selectively reacts with the target gas through a specific electrode material to generate an electric signal proportional to the concentration. The carbon monoxide sensor uses a nano-scale metal oxide semiconductor material and measures the concentration through the resistance change caused by gas adsorption, featuring fast response and high sensitivity.

[0092] This embodiment constructs an accurate positioning system. The Beidou navigation module supports multi-mode and multi-frequency reception, and improves the positioning accuracy by combining differential positioning technology. The satellite positioning and inertial sensor data are fused through the Kalman filter algorithm to achieve continuous and stable position tracking. Meanwhile, the map matching function is realized to accurately map the positioning result to the road network.

[0093] This embodiment optimizes the data association mechanism. The spatio-temporal tag method is adopted to establish the correspondence between the monitoring data and the location information to ensure the traceability of the data. The sensor response time is considered in the data association process, and the sampling delay effect is eliminated through time alignment. Meanwhile, the data quality assessment function is realized to mark the data that does not meet the quality requirements.

[0094] This embodiment particularly focuses on the spatio-temporal consistency of the monitoring data. During the data generation process, the UTC time standard is uniformly adopted to eliminate the time zone difference. The position coordinates adopt the WGS84 coordinate system to facilitate the integration with other spatial data. Through the spatio-temporal correlation analysis of the data, the spatial distribution characteristics and the time evolution law of the pollutant concentration are identified.

[0095] Through the above technological innovations, this embodiment solves the problems such as unreasonable regional division and fixed sampling strategies in traditional environmental monitoring. In practical applications, this solution can flexibly divide the monitoring area according to the urban characteristics and achieve the optimal allocation of monitoring resources. Through the collaborative monitoring of multi-sensors and precise positioning, the quality and reliability of environmental monitoring data are significantly improved. Especially in a complex urban environment, the system can maintain stable monitoring performance and provide reliable data support for environmental management decisions.

[0096] In an embodiment of the dynamic traffic environment monitoring method of this application, it may also specifically include the following content:

[0097] Step S301: The environmental monitoring and collection device sets up a cache data pool to store the collected monitoring data. The particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration in the cache data pool are respectively mapped to the sampling location coordinates and the sampling timestamp. The environmental monitoring and collection device reads the monitoring data from the cache data pool according to a preset data cache threshold, performs standardization processing and encoding compression on the read monitoring data, and generates the monitoring data packet.

[0098] Step S302: The environmental monitoring and collection device encapsulates the format of the monitoring data packet based on a preset data transmission protocol, selects a transmission channel according to the wireless network signal quality. The environmental monitoring and collection device uploads the encapsulated monitoring data packet to the cloud computing platform through the transmission channel. The cloud computing platform parses and performs integrity verification on the received monitoring data packet, and stores the verified monitoring data in the database.

[0099] Optionally, this embodiment innovatively designs a data cache management mechanism. The environmental monitoring and collection device adopts a hierarchical data pool structure, including a real-time data layer and a persistent storage layer. The real-time data layer uses a circular buffer design, supporting high-speed data writing and reading. The persistent storage layer is implemented using flash memory to ensure that data is not lost in case of power failure. The data pool capacity is dynamically adjusted and adaptively changes according to the sampling frequency and network status.

[0100] This embodiment implements an efficient data mapping mechanism. The mapping of monitoring data to spatio-temporal information adopts a key-value pair storage structure. The key includes the timestamp and the location hash value, and the value includes the corresponding pollutant concentration data. The mapping table uses a B+ tree index to support fast spatio-temporal range queries. To improve query efficiency, the system establishes an in-memory index for hot data, and cold data is stored in flash memory.

[0101] This embodiment optimizes the data reading strategy. The data cache threshold is dynamically adjusted according to multiple factors, and the calculation formula is:

[0102] T = Tbase (α M +β N + γ Q)

[0103] Where Tbase is the baseline threshold, M is the memory usage rate, N is the network transmission rate, Q is the data quality factor, and α, β, γ are weight coefficients. This adaptive mechanism ensures the real-time and reliability of data reading.

[0104] This embodiment designs an intelligent data standardization process. Appropriate standardization methods are adopted for different sensor data, considering the range and accuracy requirements. The logarithmic transformation is used to eliminate data skewness for particulate matter concentration, and the maximum-minimum normalization is used for gas concentration data. The original accuracy information is retained during the standardization process to ensure data reducibility.

[0105] This embodiment innovatively implements a data compression algorithm. The combination of differential coding and run-length coding is adopted, and it is optimized according to the time-series characteristics of environmental data. High compression ratio coding is used for data segments with slow changes, while more detailed information is retained for data segments with sudden changes. The coding efficiency is improved through a data prediction model during the compression process.

[0106] This embodiment constructs a reliable data transmission mechanism. The packet encapsulation adopts a hierarchical protocol structure, including a data layer, a control layer, and a session layer. The data layer performs payload encapsulation, the control layer adds checksum and timing information, and the session layer is responsible for connection management. The transmission protocol supports data retransmission and flow control mechanisms to ensure reliable data transmission.

[0107] This embodiment optimizes the transmission channel selection strategy. The system monitors the signal quality indicators of each wireless channel in real time, including signal strength, bit error rate, and delay, etc. The channel quality is comprehensively evaluated through a weighted scoring method to select the optimal transmission path. At the same time, a channel switching mechanism is implemented to quickly switch to the backup channel when the current channel quality deteriorates.

[0108] This embodiment implements a complete data verification mechanism. The cloud platform adopts a multiple verification strategy, including packet header verification, CRC verification, and sequence number verification. Packets that fail the verification are required to be retransmitted. The data that passes the verification is decompressed and deserialized to restore the original data format. The system also establishes a data integrity log to record all verification exceptions.

[0109] This embodiment designs a high-performance data storage structure. A distributed time-series database is used to store monitoring data, supporting high-concurrency writing and fast querying. The database adopts a sharding storage strategy to partition data according to time and space dimensions. At the same time, a hot data caching mechanism is implemented to improve the query response speed.

[0110] Through the above technological innovations, this embodiment effectively solves problems such as data loss and unstable transmission in environmental monitoring data management. In practical applications, this solution can ensure the integrity and reliability of monitoring data, supporting real-time collection and transmission of large-scale data. Especially in the case of unstable network conditions, the system significantly improves the data transmission success rate through multi-level caching and intelligent transmission control.

[0111] The innovation points of this embodiment are mainly reflected in aspects such as data cache management, compressed transmission, and verification mechanisms. Through a series of optimized designs, the efficient management and reliable transmission of environmental monitoring data are achieved. This solution provides a strong guarantee for the stable operation of the environmental monitoring system and is of great significance for improving the environmental monitoring ability.

[0112] In an embodiment of the dynamic traffic environment monitoring method of this application, the following content may also be specifically included:

[0113] Step S401: The cloud computing platform respectively normalizes the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration in the monitoring data packet, calculates the data similarity of adjacent sampling points using the time series clustering algorithm, marks the data points with a similarity deviation greater than the preset threshold as outliers, and the cloud computing platform corrects the marked outliers based on the linear interpolation method, and uses the corrected monitoring data as the effective monitoring data;

[0114] Step S402: The cloud computing platform arranges the effective monitoring data in time series according to the sampling timestamp, maps the effective monitoring data within the same time period to a two-dimensional spatial grid based on the sampling position coordinates, uses the Kriging interpolation method to supplement the missing measurement points in the two-dimensional spatial grid, constructs the supplemented grid data into the spatio-temporal data matrix, where the row vector of the spatio-temporal data matrix represents the sampling time and the column vector represents the spatial position.

[0115] Optionally, this embodiment realizes intelligent data normalization processing. For the dimensional differences and numerical distribution characteristics of different pollutants, an adaptive normalization method is adopted. The normalization formula is:

[0116] X' = (X - μ) / (σ k)

[0117] Where X is the original data, μ is the mean value within the sliding time window, σ is the standard deviation, and k is the adjustment coefficient. This method can maintain the relative change trend of the data, eliminate the influence of dimensions at the same time, and facilitate subsequent analysis and processing.

[0118] This embodiment innovatively designs a time series clustering analysis method. An improved DBSCAN algorithm is used to cluster the time series data, and outliers are adaptively identified through dynamic density thresholds. The similarity of adjacent sampling points is calculated based on multi-dimensional features, including numerical differences, change trends, and spatio-temporal correlations. The algorithm identifies outliers through density connectivity analysis, effectively avoiding the limitations of traditional fixed threshold methods.

[0119] This embodiment optimizes the outlier correction strategy. For the marked outliers, the system first analyzes the causes of outliers and differentiates between sudden pollution events and equipment failures. For the outliers caused by equipment failures, the local linear interpolation method is used for correction. The interpolation process takes into account the continuity of the time series and spatial correlation to ensure the rationality of the correction results.

[0120] This embodiment realizes efficient data time-series processing. The effective monitoring data is constructed into an ordered sequence according to timestamps, and a B+ tree index is used to support fast time range queries. The time consistency of the data is maintained during the serialization process, and the data with different sampling frequencies is time-aligned to ensure the comparability of the data.

[0121] This embodiment designs an innovative space mapping mechanism. Based on the sampling position coordinates, the monitoring area is divided into equally spaced grids, and the grid size is dynamically adjusted according to the monitoring density. The data of irregularly distributed sampling points is mapped to regular grids through spatial scatter interpolation to establish a unified spatial reference system.

[0122] This embodiment constructs an intelligent Kriging interpolation model. Aiming at the spatial autocorrelation characteristics of environmental data, the ordinary Kriging method is used for interpolation calculation. The variogram model is obtained through training with historical data and can accurately describe the spatial continuity of pollutants. The interpolation process takes into account the influence of environmental factors such as terrain and buildings to improve the interpolation accuracy.

[0123] This embodiment optimizes the data supplementation strategy. For the missing measurement points in the grid, the system first analyzes the distribution characteristics of the surrounding valid data. The interpolation search radius is determined according to the spatial correlation, and the value of the missing measurement point is estimated by Kriging interpolation. The reliability of the interpolation result is evaluated through cross-validation, and the areas with high uncertainty are marked.

[0124] This embodiment innovatively constructs a spatio-temporal data matrix. The matrix adopts a sparse storage structure, effectively reducing the storage overhead. The row vectors represent the monitoring data at different time points, and the column vectors represent the observed values at different spatial positions. The matrix elements contain pollutant concentration values and data reliability indicators, supporting subsequent spatio-temporal analysis.

[0125] This embodiment realizes a complete data quality control system. Quality inspection points are set at each link from data normalization to matrix construction. Through methods such as data consistency inspection, spatial autocorrelation analysis, and time continuity evaluation, the reliability of the processing results is ensured. At the same time, a data processing log is established to record abnormal situations and processing processes.

[0126] Through the above technological innovations, this embodiment effectively solves problems such as outlier interference and uneven spatial coverage in environmental monitoring data processing. In practical applications, this solution can accurately identify and correct abnormal data, and achieve complete coverage of monitoring data through spatial interpolation. Especially in the case of unevenly distributed sampling points, the system significantly improves the spatial resolution of environmental monitoring through an intelligent interpolation algorithm.

[0127] The main innovation points of this embodiment are reflected in data normalization processing, outlier identification, and spatial interpolation, etc. Through a series of data processing optimizations, the standardization and gridification of environmental monitoring data are achieved, providing a high-quality data basis for subsequent pollutant diffusion analysis. This solution has important practical value in improving the quality and spatial coverage of environmental monitoring data.

[0128] In an embodiment of the dynamic traffic environment monitoring method of this application, the following content may also be specifically included:

[0129] Step S501: The cloud computing platform inputs the spatio-temporal data matrix into a multi-channel convolutional neural network. The multi-channel convolutional neural network uses a two-channel structure to extract features in the time dimension and the space dimension respectively. The time dimension features are extracted by a one-dimensional convolutional layer to obtain the temporal variation pattern of pollutant concentration, and the space dimension features are extracted by a two-dimensional convolutional layer to obtain the spatial distribution law of pollutant concentration. The time dimension features and the space dimension features are fused to obtain pollutant spatio-temporal features, and a pollutant concentration distribution model is constructed based on the pollutant spatio-temporal features;

[0130] Step S502: The cloud computing platform reads historical monitoring data from the database, clusters the historical monitoring data according to pollutant types and concentration change trends, and uses a long short-term memory network to establish the pollutant migration model. The pollutant migration model takes the output features of the pollutant concentration distribution model as input, combines meteorological data to predict the diffusion trend of pollutants at different spatio-temporal scales, and generates a spatio-temporal evolution sequence of the pollutant concentration field based on the diffusion trend.

[0131] Optionally, this embodiment innovatively designs a multi-channel convolutional neural network architecture. The network uses a two-channel structure to separately process feature extraction in the time and space dimensions. The time channel uses a one-dimensional convolutional structure, and the size of the convolutional kernel is dynamically adjusted according to the time scale to capture change patterns of different periods. The space channel uses a two-dimensional convolutional layer, and local and global spatial features are extracted through a multi-scale feature pyramid.

[0132] This embodiment optimizes the temporal feature extraction mechanism. The one-dimensional convolutional layer adopts a multi-scale convolutional kernel design, and the mathematical expression for temporal feature extraction is:

[0133] Ft = Conv1D(X, W) A

[0134] where X is the input time series data, W is the convolutional kernel weight, and A is the attention weight matrix. The feature weights at different time scales are adaptively adjusted through the attention mechanism to highlight the key time series patterns. Residual connections are used between convolutional layers to alleviate the problem of vanishing gradients in the training of deep networks.

[0135] This embodiment realizes efficient spatial feature extraction. The two-dimensional convolutional layer adopts a dilated convolution design to expand the receptive field without increasing the number of parameters. By adjusting the dilation rate, the adaptive extraction of multi-scale spatial features is achieved. At the same time, a spatial attention mechanism is introduced to enhance the feature extraction ability for high-concentration regions of pollutants.

[0136] This embodiment innovatively designs a feature fusion strategy. The spatio-temporal feature fusion adopts an adaptive weighting method, and the fusion weights are automatically learned through backpropagation. The fusion process takes into account the correlation between time and space features and controls the flow of features through a gating mechanism. The fused features retain both the time series change information and the spatial distribution law.

[0137] This embodiment optimizes the historical data processing flow. An improved K-means algorithm is used to cluster the historical data, and the clustering features include pollutant concentration levels, change rates, and periodicity, etc. By dynamically updating the clustering centers, it adapts to the time-varying characteristics of pollutant distribution. The clustering results are used to guide subsequent model training to improve the prediction accuracy.

[0138] This embodiment constructs an innovative LSTM transfer model. The model adopts a hierarchical structure, including an encoder and a decoder. The encoder encodes the pollutant concentration distribution features and meteorological data into hidden state vectors, and the decoder generates prediction results at different spatio-temporal scales based on the hidden states. The model processes long sequence dependencies through the attention mechanism to improve the prediction accuracy.

[0139] This embodiment realizes multi-source data fusion prediction. The fusion of meteorological data and pollutant characteristics adopts an early fusion strategy, and feature combination is performed at the input layer of the LSTM. The model inputs include meteorological elements such as wind direction, wind speed, temperature, and humidity, and a non-linear transformation is used to establish the correlation between meteorological conditions and pollutant migration.

[0140] This embodiment designs a diffusion trend prediction mechanism. The model is based on a method that combines physical diffusion laws and data-driven approaches to predict the migration and diffusion process of pollutants at different spatio-temporal scales. The prediction results include the concentration field distribution and diffusion rate, which can be used to guide pollution prevention and control decisions.

[0141] This embodiment optimizes the spatio-temporal expression of the prediction results. The spatio-temporal evolution sequence of the concentration field is represented by a four-dimensional tensor, which includes three dimensions: time, space, and pollutant type. The sequence generation process takes into account the uncertainty of the prediction, and describes the credibility of the prediction results through probability distributions. At the same time, the visual expression of the prediction results is realized, intuitively showing the pollutant diffusion process.

[0142] Through the above technological innovations, this embodiment effectively solves the problems of insufficient spatio-temporal feature extraction and low prediction accuracy in traditional pollutant prediction models. In practical applications, this solution can accurately capture the spatio-temporal variation law of pollutant concentration and achieve multi-scale prediction. Especially under complex meteorological conditions, the system significantly improves the accuracy and reliability of the prediction through multi-source data fusion and deep learning models.

[0143] The innovation points of this embodiment are mainly reflected in aspects such as network structure design, feature extraction and fusion, and prediction model construction. Through deep learning technology, accurate modeling and prediction of pollutant concentration distribution are realized, providing a scientific basis for urban air quality management. This solution has important application value in the fields of environmental monitoring and early warning.

[0144] In an embodiment of the dynamic traffic environment monitoring method of this application, it may also specifically include the following content:

[0145] Step S601: The cloud computing platform calculates the pollutant concentration gradient and change rate of each region based on the pollutant diffusion trend prediction results, inputs the pollutant concentration gradient and the change rate into a fuzzy neural network, and the fuzzy neural network divides the monitoring region into risk levels according to a preset pollutant concentration threshold, generating a pollution risk distribution map including high, medium, and low three risk levels. The cloud computing platform performs grid processing on the pollution risk distribution map to obtain the pollution risk level region;

[0146] Step S602: The cloud computing platform constructs the monitoring resource scheduling model using a deep reinforcement learning algorithm. The monitoring resource scheduling model takes the pollution risk level region as the state space, takes the real-time position information and remaining power of the power-assisted bicycle as constraint conditions, calculates the optimal monitoring path based on the value function iteration method, and the monitoring resource scheduling model assigns inspection routes and sampling frequencies to each power-assisted bicycle according to the optimal monitoring path, generating the monitoring task allocation plan including the sampling region and sampling duration.

[0147] Optionally, this embodiment innovatively designs a pollutant risk assessment mechanism. The cloud computing platform calculates the spatial change of pollutant concentration using a gradient operator based on the diffusion trend prediction results. The comprehensive risk assessment value of the concentration gradient G and the change rate R is calculated by the following formula:

[0148] Risk = w1 |∇G| + w2 |dR / dt|

[0149] where ∇G is the concentration gradient vector, dR / dt is the concentration change rate, and w1 and w2 are dynamic weight coefficients, which are adjusted according to the hazard characteristics of different pollutants.

[0150] This embodiment realizes an intelligent fuzzy neural network structure. The network adopts a five-layer architecture, including an input layer, a fuzzification layer, a rule layer, a defuzzification layer, and an output layer. The input features are transformed into linguistic variables through fuzzification processing, and the risk level is evaluated through adaptive fuzzy rule inference. The network training uses the backpropagation algorithm to dynamically adjust the membership function parameters and rule weights.

[0151] This embodiment optimizes the risk level classification strategy. The pollutant concentration threshold is determined according to historical monitoring data and environmental standards, and the monitoring area is divided into high, medium, and low risk levels in combination with the fuzzy inference results. The high-risk area corresponds to the area where the pollutant concentration continuously exceeds the standard or rises rapidly, the medium-risk area indicates the area where the concentration is close to the threshold, and the low-risk area is the area where the concentration is stable within the safe range.

[0152] This embodiment constructs an accurate grid processing method. The risk distribution map uses adaptive grid division, and a finer grid resolution is used in the high-risk area. The grid process considers geographical boundaries and building distributions to ensure the practicality of risk area division. At the same time, a dynamic update mechanism for grid data is implemented to adjust the risk level in real time according to the pollutant diffusion trend.

[0153] This embodiment innovatively designs a deep reinforcement learning scheduling model. The model adopts the Actor-Critic architecture. The Actor network is responsible for generating scheduling strategies, and the Critic network evaluates the strategy value. The state space includes information such as the risk level distribution and the location of monitoring devices, and the action space includes path selection and sampling configuration. The design of the reward function considers multiple objectives such as monitoring coverage, risk responsiveness, and resource efficiency.

[0154] This embodiment optimizes the constraint handling mechanism. The position information and battery status of the assistive bicycle are converted into constraints, and the constraints are integrated into the reinforcement learning framework through the penalty function method. The battery constraint ensures that the planned inspection route can be completed within the energy range of the device, and the position constraint considers the mobility and accessibility of the device.

[0155] This embodiment implements an efficient value function iteration algorithm. By adopting a prioritized experience replay mechanism, it focuses on learning the value estimation of key state transitions. The design of a double Q-network reduces the bias in value function estimation and improves the learning efficiency. During the iteration process, the exploration rate is dynamically adjusted to balance the relationship between exploration and exploitation.

[0156] This embodiment designs an intelligent path planning strategy. Based on the output of a deep reinforcement learning model, it combines the A* algorithm to plan specific inspection routes. The path planning takes into account road network constraints and traffic conditions to ensure the feasibility of the planned routes. Meanwhile, it realizes the function of dynamic obstacle avoidance and can adjust the routes according to real-time road conditions.

[0157] This embodiment optimizes the method for generating sampling strategies. The system adaptively adjusts the sampling frequency according to the risk level and the distribution of monitoring devices. A higher sampling frequency is adopted in high-risk areas to ensure timely detection of pollution changes. The sampling duration is determined according to the characteristics of pollutants and the daily variation law to ensure the representativeness of the data.

[0158] This embodiment constructs a complete task allocation mechanism. The monitoring task allocation plan includes elements such as inspection routes, sampling point locations, sampling frequencies, and monitoring durations. The system ensures that high-risk areas are preferentially covered by monitoring through task priority sorting. Meanwhile, it realizes a task conflict detection and coordination mechanism to avoid resource competition.

[0159] Through the above technological innovations, this embodiment effectively solves problems such as unreasonable resource scheduling and lagged response in traditional environmental monitoring. In practical applications, this solution can intelligently schedule monitoring resources according to the pollution risk distribution and achieve precise allocation of monitoring forces. Especially in the response to sudden pollution events, the system quickly generates an optimized monitoring plan through a deep reinforcement learning algorithm, significantly improving the efficiency and pertinence of environmental monitoring.

[0160] The innovation points of this embodiment are mainly reflected in aspects such as risk assessment models, resource scheduling algorithms, and task planning strategies. It realizes the intelligent scheduling of environmental monitoring resources through artificial intelligence technology, providing strong support for urban environmental governance. This solution has important practical value in improving the efficiency of environmental monitoring and emergency response capabilities.

[0161] In an embodiment of the dynamic traffic environment monitoring method of this application, it may also specifically include the following content:

[0162] Step S701: The cloud computing platform encodes the monitoring task allocation scheme according to the data transmission protocol, and sends the encoded monitoring task allocation scheme to each power-assisted bicycle based on the wireless communication network. After receiving the monitoring task allocation scheme, the power-assisted bicycle decodes it, plans the sampling trajectory according to the inspection route, and the cloud computing platform tracks the task execution status of each power-assisted bicycle in real time. When it detects that the power-assisted bicycle deviates from the inspection route, it recalculates the optimal monitoring path.

[0163] Step S702: The cloud computing platform divides the road network in the monitoring area into grids based on the pollutant diffusion trend prediction result, calculates the pollutant concentration change trend in each grid unit in the future time period, compares the pollutant concentration change trend with the preset pollutant concentration threshold to obtain the pollution risk index, determines the traffic flow regulation area according to the spatial distribution of the pollution risk index, and generates the warning information including the vehicle flow guidance scheme based on the correlation model between traffic flow and pollutant concentration.

[0164] Optionally, this embodiment designs an intelligent task allocation protocol. The monitoring task allocation scheme adopts a hierarchical coding structure, including a basic configuration layer and an execution instruction layer. The basic configuration layer defines device parameters and sampling requirements, and the execution instruction layer includes the inspection route and sampling strategy. The coding adopts a binary compression format, and the storage space is optimized through bitmap indexing. The reliability of the task coding is ensured by the redundant check code.

[0165] This embodiment realizes an efficient task distribution mechanism. The cloud platform selects the optimal transmission path according to the network condition, and distributes tasks in a push-pull combination manner. Emergency tasks are immediately distributed by the push method, and regular tasks allow devices to pull them regularly. The transmission process adopts a fragmentation mechanism to segment and transmit large task schemes to ensure the reliability of transmission.

[0166] This embodiment optimizes the task decoding process. After receiving the task, the power-assisted bicycle first performs integrity verification, and then parses the task configuration parameters. The decoding process adopts a state machine design to ensure the accuracy of parameter parsing. After decoding is completed, the task parameters are loaded into the execution queue, and a confirmation message is returned to the cloud platform.

[0167] This embodiment innovatively designs a path planning algorithm. Based on A The algorithm implements sampling trajectory planning, and the path weight W is calculated by the following formula:

[0168] W = αD +β P + γE

[0169] Where D is the path distance, P is the predicted pollutant concentration, E is the remaining battery power, and α, β, and γ are weight coefficients. This multi-objective optimization method ensures the rationality of the sampling path.

[0170] This embodiment realizes a real-time task tracking mechanism. The cloud platform continuously monitors the vehicle position through GPS positioning and uses the Kalman filter algorithm to smooth the trajectory data. When it detects that the vehicle deviates from the predetermined route, the system evaluates the reason for the deviation and distinguishes between active obstacle avoidance and passive deviation. For passive deviation situations, it triggers the path replanning process.

[0171] This embodiment constructs an innovative grid processing method. It divides the grid based on the road network topology structure, and the grid size is dynamically adjusted according to the road density. Each grid unit contains road attributes and pollutant prediction data. It analyzes the connectivity between grids through graph theory algorithms and establishes a pollutant transmission network model.

[0172] This embodiment optimizes the risk index calculation strategy. The pollution risk index R adopts a multi-factor comprehensive evaluation method:

[0173] R = w1 C + w2V + w3 T

[0174] Where C is the predicted concentration value, V is the change rate, T is the duration, and w1, w2, and w3 are dynamic weights. This method comprehensively considers the hazard characteristics of pollutants.

[0175] This embodiment designs an intelligent traffic control mechanism. Based on the spatial distribution of the risk index, it uses a clustering algorithm to identify high-risk areas. The system analyzes the road network structure to determine key control nodes. Through the coupling model of traffic flow and pollutant concentration, it evaluates the effects of different control strategies.

[0176] This embodiment realizes an accurate vehicle flow guidance scheme. The guidance scheme considers the road network capacity and pollutant diffusion characteristics and uses the dynamic programming algorithm to calculate the optimal diversion path. The scheme includes multi-level control measures, from recommended detours to mandatory traffic restrictions, which are gradually upgraded according to the risk level. At the same time, it establishes an effect evaluation mechanism to dynamically adjust the control strategy.

[0177] This embodiment innovatively designs a warning information generation mechanism. The warning information adopts a hierarchical and classified structure, and customizes the information content for different user groups. The information includes pollution risk tips, detour suggestions, and control measures, etc. The system publishes warnings through multiple channels to ensure the timely delivery of information.

[0178] Through the above technological innovations, this embodiment effectively solves the problems of low task execution efficiency and lagging warning response in traditional environmental monitoring. In practical applications, this solution can achieve intelligent allocation and dynamic adjustment of monitoring tasks, and reduce the risk of pollutant accumulation through precise traffic control. Especially in the response to heavy pollution weather, the system significantly improves the effect of early warning and prevention and control through multi-dimensional data analysis and intelligent decision-making.

[0179] The innovation points of this embodiment are mainly reflected in aspects such as task management, path planning, and traffic regulation. Through deep learning and optimization algorithms, the collaborative linkage between environmental monitoring and traffic management is realized. This solution provides technical support for improving urban air quality and has important practical application value. This system plays an active role in enhancing the urban environmental monitoring ability and improving air quality, and provides a useful exploration for the construction of smart cities.

[0180] In order to effectively solve the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improve the accuracy and practical value of traffic environment monitoring, this application provides an embodiment of a dynamic traffic environment monitoring device for implementing all or part of the content of the dynamic traffic environment monitoring method. See Figure 2 , the dynamic traffic environment monitoring device specifically includes the following content:

[0181] An environmental acquisition module 10, configured to configure an assisted bicycle acquisition network with environmental monitoring acquisition devices. The environmental monitoring acquisition devices acquire particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, record the sampling position coordinates and sampling timestamps, and the environmental monitoring acquisition devices organize the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into monitoring data packets and upload them to the cloud computing platform;

[0182] A pollution prediction module 20, configured to perform data preprocessing on the received monitoring data packets by the cloud computing platform, identify outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on a time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix according to the sampling position coordinates and the sampling timestamps for the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, establishes a pollutant concentration distribution model, and the cloud computing platform trains a pollutant migration model in combination with historical monitoring data, and inputs the output features of the pollutant concentration distribution model into the pollutant migration model to generate a pollutant diffusion trend prediction result;

[0183] The dynamic monitoring module 30 is used for the cloud computing platform to divide the pollution risk level areas according to the pollutant diffusion trend prediction result, construct a monitoring resource scheduling model based on the deep reinforcement learning algorithm, use the pollution risk level areas, the real-time position information and the remaining power of the power-assisted bicycle as input parameters to generate a monitoring task allocation plan, the cloud computing platform sends the monitoring task allocation plan including the inspection route to the power-assisted bicycle, calculates the pollution risk index of each road section in the future time period according to the pollutant diffusion trend prediction result, and generates a warning information including traffic control strategies.

[0184] As can be seen from the above description, the dynamic traffic environment monitoring device provided by the embodiments of the present application can realize the mobile sampling of pollutants such as particulate matter, nitrogen oxides and carbon monoxide by constructing a power-assisted bicycle environment monitoring and acquisition network. The cloud computing platform performs outlier identification based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with the pollutant migration model. The deep reinforcement learning algorithm is used to realize the intelligent scheduling of monitoring resources, and an optimized inspection plan is generated according to the pollution risk level, vehicle position and power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0185] At the hardware level, in order to effectively solve the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis and resource scheduling, and significantly improve the accuracy and practical value of traffic environment monitoring, the present application provides an embodiment of an electronic device for implementing all or part of the content in the dynamic traffic environment monitoring method. The electronic device specifically includes the following contents:

[0186] A processor, a memory, a communication interface and a bus; wherein, the processor, the memory and the communication interface complete mutual communication through the bus; the communication interface is used to realize the information transmission between the dynamic traffic environment monitoring device and related devices such as the core business system, the user terminal and the relevant database, etc. The logic controller can be a desktop computer, a tablet computer, a mobile terminal, etc., and this embodiment is not limited thereto. In this embodiment, the logic controller can be implemented with reference to the embodiments of the dynamic traffic environment monitoring method and the embodiments of the dynamic traffic environment monitoring device in the embodiments, and the content thereof is incorporated herein, and the repeated parts will not be described again.

[0187] It can be understood that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0188] In practical applications, part of the dynamic traffic environment monitoring method can be executed on the side of the electronic device as described above, or all operations can be completed in the client device. Specifically, it can be selected according to the processing capacity of the client device and the limitations of the user usage scenario, etc. This application does not make any limitations in this regard. If all operations are completed in the client device, the client device may further include a processor.

[0189] The above-mentioned client device may have a communication module (i.e., a communication unit), and can communicate with a remote server to realize data transmission with the server. The server may include a server on the side of the task scheduling center, and in other implementation scenarios, it may also include a server of an intermediate platform, such as a server of a third-party server platform having a communication link with the task scheduling center server. The server may include a single computer device, or may include a server cluster composed of multiple servers, or a server structure of a distributed device.

[0190] Figure 3 It is a schematic block diagram of the system composition of the electronic device 9600 according to an embodiment of the present application. As Figure 3 shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It should be noted that this Figure 3 is exemplary; other types of structures can also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0191] In one embodiment, the function of the dynamic traffic environment monitoring method can be integrated into the central processing unit 9100. Among them, the central processing unit 9100 may be configured to perform the following controls:

[0192] Step S101: Configure an assisted bicycle collection network with an environmental monitoring collection device. The environmental monitoring collection device collects the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling position coordinates and sampling timestamps, and the environmental monitoring collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into a monitoring data packet and uploads it to the cloud computing platform;

[0193] Step S102: The cloud computing platform performs data preprocessing on the received monitoring data packets, identifies outliers for the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix with the effective monitoring data according to the sampling position coordinates and the sampling timestamps, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform trains a pollutant migration model in combination with historical monitoring data, inputs the output features of the pollutant concentration distribution model into the pollutant migration model, and generates a pollutant diffusion trend prediction result;

[0194] Step S103: The cloud computing platform divides the pollution risk level areas according to the pollutant diffusion trend prediction result, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, uses the pollution risk level areas, the real-time position information and the remaining power of the power-assisted bicycle as input parameters to generate a monitoring task allocation plan. The cloud computing platform sends the monitoring task allocation plan including the inspection route to the power-assisted bicycle, calculates the pollution risk index of each road section in the future period according to the pollutant diffusion trend prediction result, and generates a warning information including traffic control strategies.

[0195] As can be seen from the above description, the electronic device provided in the embodiment of the present application realizes the mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide by constructing a power-assisted bicycle environmental monitoring and acquisition network. The cloud computing platform performs outlier identification based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. The deep reinforcement learning algorithm is used to realize the intelligent scheduling of monitoring resources, and an optimized inspection plan is generated according to the pollution risk level, vehicle position, and power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0196] In another embodiment, the dynamic traffic environment monitoring device can be separately configured from the central processing unit 9100. For example, the dynamic traffic environment monitoring device can be configured as a chip connected to the central processing unit 9100, and the functions of the dynamic traffic environment monitoring method are realized through the control of the central processing unit.

[0197] As Figure 3 shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It should be noted that the electronic device 9600 does not necessarily have to include all the components shown in Figure 3 ; in addition, the electronic device 9600 may further includeFigure 3 For components not shown herein, reference may be made to the prior art.

[0198] As Figure 3 shown, the central processing unit 9100, sometimes also referred to as a controller or operation control, may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives inputs and controls the operation of the various components of the electronic device 9600.

[0199] Among them, the memory 9140 can be, for example, one or more of a buffer, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory, or other suitable devices. It can store the above-mentioned information related to failures, and can also store programs for executing relevant information. And the central processing unit 9100 can execute the programs stored in the memory 9140 to achieve information storage or processing, etc.

[0200] The input unit 9120 provides inputs to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to supply power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display can be, for example, an LCD display, but is not limited thereto.

[0201] The memory 9140 can be a solid-state memory. For example, it can be a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It can also be such a memory that stores information even when power is off, can be selectively erased and has more data. Examples of such a memory are sometimes referred to as EPROMs, etc. The memory 9140 can also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 can include an application / function storage section 9142, which is used to store application programs and function programs or the processes for operating the electronic device 9600 through the central processing unit 9100.

[0202] The memory 9140 can also include a data storage section 9143, which is used to store data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage section 9144 of the memory 9140 can include various drivers of the electronic device for communication functions and / or for executing other functions of the electronic device (such as a messaging application, an address book application, etc.).

[0203] The communication module 9110 is a transmitter / receiver that transmits and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which can be the same as in the case of a conventional mobile communication terminal.

[0204] Based on different communication technologies, multiple communication modules 9110, such as a cellular network module, a Bluetooth module, and / or a wireless local area network module, etc., can be provided in the same electronic device. The communication module 9110 (transmitter / receiver) is also coupled to the speaker 9131 and the microphone 9132 via the audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby implementing the usual telecommunication functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local device through the microphone 9132, and the sound stored on the local device can be played through the speaker 9131.

[0205] Embodiments of the present application also provide a computer-readable storage medium capable of implementing all steps in the dynamic traffic environment monitoring method where the execution subject in the above embodiments is a server or a client. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all steps in the dynamic traffic environment monitoring method where the execution subject in the above embodiments is a server or a client are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0206] Step S101: Configure an assisted bicycle collection network with an environmental monitoring collection device. The environmental monitoring collection device collects the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling location coordinates and the sampling timestamp. The environmental monitoring collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling location coordinates, and the sampling timestamp into a monitoring data packet and uploads it to the cloud computing platform;

[0207] Step S102: The cloud computing platform preprocesses the received monitoring data packets, identifies outliers for the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix based on the sampling location coordinates and the sampling timestamps of the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform trains a pollutant migration model in combination with historical monitoring data, inputs the output features of the pollutant concentration distribution model into the pollutant migration model, and generates a pollutant diffusion trend prediction result;

[0208] Step S103: The cloud computing platform divides pollution risk level regions according to the pollutant diffusion trend prediction result, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, uses the pollution risk level regions, the real-time location information and the remaining power of the power-assisted bicycle as input parameters to generate a monitoring task allocation plan. The cloud computing platform sends the monitoring task allocation plan including the inspection route to the power-assisted bicycle, calculates the pollution risk index of each road section in the future period according to the pollutant diffusion trend prediction result, and generates a warning message including traffic control strategies.

[0209] As can be seen from the above description, the computer-readable storage medium provided by the embodiments of the present application realizes mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide by constructing a power-assisted bicycle environmental monitoring and collection network. The cloud computing platform performs outlier identification based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. The deep reinforcement learning algorithm is used to realize the intelligent scheduling of monitoring resources, and an optimized inspection plan is generated according to the pollution risk level, vehicle location, and power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0210] The embodiments of the present application also provide a computer program product capable of implementing all the steps in the dynamic traffic environment monitoring method with the execution subject being a server or a client in the above embodiments. When the computer program / instructions are executed by a processor, the steps of the dynamic traffic environment monitoring method are implemented. For example, the computer program / instructions implement the following steps:

[0211] Step S101: Configure an assisted bicycle collection network with an environmental monitoring and collection device. The environmental monitoring and collection device collects particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling location coordinates and sampling timestamps. The environmental monitoring and collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling location coordinates, and the sampling timestamps into a monitoring data packet and uploads it to the cloud computing platform;

[0212] Step S102: The cloud computing platform performs data preprocessing on the received monitoring data packet, identifies outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix according to the sampling location coordinates and the sampling timestamps for the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform trains a pollutant migration model in combination with historical monitoring data, and inputs the output features of the pollutant concentration distribution model into the pollutant migration model to generate a pollutant diffusion trend prediction result;

[0213] Step S103: The cloud computing platform divides pollution risk level regions according to the pollutant diffusion trend prediction result, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, uses the pollution risk level regions, the real-time location information and remaining battery power of the assisted bicycle as input parameters to generate a monitoring task allocation plan. The cloud computing platform sends the monitoring task allocation plan including the inspection route to the assisted bicycle, calculates the pollution risk index for each road section in the future time period according to the pollutant diffusion trend prediction result, and generates a warning message including traffic control strategies.

[0214] As can be seen from the above description, the computer program product provided by the embodiments of the present application realizes mobile sampling of pollutants such as particulate matter, nitrogen oxides, and carbon monoxide by constructing an assisted bicycle environmental monitoring and collection network. The cloud computing platform performs outlier identification based on time series clustering, constructs a pollutant concentration distribution model through a multi-channel convolutional neural network, and generates a diffusion trend prediction in combination with a pollutant migration model. The deep reinforcement learning algorithm is used to realize intelligent scheduling of monitoring resources, and an optimized inspection plan is generated according to the pollution risk level, vehicle location, and battery power. This method effectively solves the deficiencies of traditional technologies in aspects such as monitoring coverage, data analysis, and resource scheduling, and significantly improves the accuracy and practical value of traffic environment monitoring.

[0215] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, apparatus, or computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0216] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (apparatuses), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0217] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0218] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0219] Specific embodiments are applied in the present invention to elaborate on the principles and implementation manners of the present invention. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for monitoring a dynamic traffic environment, characterized in that, The method includes: Configuring an assisted bicycle collection network with an environmental monitoring and collection device, where the environmental monitoring and collection device collects particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling position coordinates and sampling timestamps, and the environmental monitoring and collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into a monitoring data packet and uploads it to the cloud computing platform; The cloud computing platform performs data preprocessing on the received monitoring data packet, identifies outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix according to the sampling position coordinates and the sampling timestamps for the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi - convolutional neural network, and establishes a pollutant concentration distribution model. The cloud computing platform combines historical monitoring data to train a pollutant migration model, inputs the output features of the pollutant concentration distribution model into the pollutant migration model, and generates a pollutant diffusion trend prediction result; The cloud computing platform divides pollution risk level regions according to the pollutant diffusion trend prediction result, constructs a monitoring resource scheduling model based on the deep reinforcement learning algorithm, uses the pollution risk level regions, the real - time position information and remaining battery power of the assisted bicycles as input parameters to generate a monitoring task allocation plan. The cloud computing platform sends the monitoring task allocation plan including the inspection route to the assisted bicycles, calculates the pollution risk index of each road section in the future time period according to the pollutant diffusion trend prediction result, and generates a warning message including traffic control strategies.

2. The dynamic traffic environment monitoring method according to claim 1, wherein The configuration of the assisted bicycle collection network with an environmental monitoring and collection device, where the environmental monitoring and collection device collects particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling position coordinates and sampling timestamps, includes: Dividing urban roads into regions based on road traffic flow distribution and wireless communication signal strength. The cloud computing platform calculates the monitoring point density threshold for each region according to traffic flow density, road type, and traffic period, determines the number of assisted bicycles required in each region based on the monitoring point density threshold, generates collection task configuration parameters including sampling period and monitoring duration, and distributes the collection task configuration parameters to the assisted bicycles; The particulate matter sensor in the environmental monitoring and collection device detects fine particulate matter in the air based on the principle of laser scattering. The nitrogen oxide sensor in the environmental monitoring and collection device obtains the nitrogen oxide content by generating an electrical signal through the reaction of electrode materials. The carbon monoxide sensor in the environmental monitoring and collection device measures the carbon monoxide concentration according to the change in the resistance of the metal oxide semiconductor material. The environmental monitoring and collection device obtains the sampling position coordinates based on the Beidou navigation and positioning module, and associates the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration with the sampling position coordinates to generate monitoring data.

3. The dynamic traffic environment monitoring method according to claim 1, characterized in that The environmental monitoring and collection device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling location coordinates, and the sampling timestamp into a monitoring data packet and uploads it to the cloud computing platform, including: The environmental monitoring and collection device sets up a cache data pool to store the collected monitoring data. The particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration in the cache data pool are respectively mapped to the sampling location coordinates and the sampling timestamp. The environmental monitoring and collection device reads the monitoring data from the cache data pool according to a preset data cache threshold, performs standardization processing and coding compression on the read monitoring data, and generates the monitoring data packet; The environmental monitoring and collection device performs format encapsulation on the monitoring data packet based on a preset data transmission protocol, selects a transmission channel according to the wireless network signal quality. The environmental monitoring and collection device uploads the encapsulated monitoring data packet to the cloud computing platform through the transmission channel. The cloud computing platform parses and performs integrity verification on the received monitoring data packet, and stores the monitoring data that passes the verification in the database.

4. The dynamic traffic environment monitoring method according to claim 1, characterized in that, The cloud computing platform performs data preprocessing on the received monitoring data packet, identifies outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on a time series clustering algorithm, and obtains effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix based on the effective monitoring data according to the sampling location coordinates and the sampling timestamp, including: The cloud computing platform respectively performs normalization processing on the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration in the monitoring data packet, calculates the data similarity of adjacent sampling points using a time series clustering algorithm, marks the data points with a similarity deviation greater than a preset threshold as outliers, and the cloud computing platform corrects the marked outliers based on a linear interpolation method, and uses the corrected monitoring data as the effective monitoring data; The cloud computing platform arranges the effective monitoring data in time series according to the sampling timestamp, maps the effective monitoring data within the same time period to a two-dimensional spatial grid based on the sampling location coordinates, uses the Kriging interpolation method to supplement the missing measurement points in the two-dimensional spatial grid, and constructs the supplemented grid data into the spatio-temporal data matrix. The row vector of the spatio-temporal data matrix represents the sampling time, and the column vector represents the spatial position.

5. The dynamic traffic environment monitoring method according to claim 1, characterized in that, The multi-channel convolutional neural network is used to extract features from the spatio-temporal data matrix to establish a pollutant concentration distribution model. The cloud computing platform combines historical monitoring data to train a pollutant migration model, and inputs the output features of the pollutant concentration distribution model into the pollutant migration model to generate a pollutant diffusion trend prediction result, including: The cloud computing platform inputs the spatio-temporal data matrix into a multi-channel convolutional neural network. The multi-channel convolutional neural network uses a two-channel structure to extract features in the time dimension and the spatial dimension respectively. The features in the time dimension are used to extract the temporal change pattern of pollutant concentration through a one-dimensional convolutional layer, and the features in the spatial dimension are used to extract the spatial distribution law of pollutant concentration through a two-dimensional convolutional layer. The features in the time dimension and the features in the spatial dimension are fused to obtain pollutant spatio-temporal features, and a pollutant concentration distribution model is constructed based on the pollutant spatio-temporal features; The cloud computing platform reads historical monitoring data from the database, clusters the historical monitoring data according to pollutant types and concentration change trends, and uses a long short-term memory network to establish the pollutant migration model. The pollutant migration model takes the output features of the pollutant concentration distribution model as input, combines meteorological data to predict the diffusion trend of pollutants at different spatio-temporal scales, and generates a spatio-temporal evolution sequence of the pollutant concentration field based on the diffusion trend.

6. The dynamic traffic environment monitoring method according to claim 1, characterized in that The cloud computing platform divides pollution risk level regions according to the pollutant diffusion trend prediction results, constructs a monitoring resource scheduling model based on a deep reinforcement learning algorithm, and uses the pollution risk level regions, the real-time position information and remaining power of the assisted bicycles as input parameters to generate a monitoring task allocation plan, including: The cloud computing platform calculates the pollutant concentration gradient and change rate of each region based on the pollutant diffusion trend prediction results, inputs the pollutant concentration gradient and the change rate into a fuzzy neural network. The fuzzy neural network divides the risk level of the monitoring region according to a preset pollutant concentration threshold, generates a pollution risk distribution map including high, medium and low three risk levels, and the cloud computing platform performs grid processing on the pollution risk distribution map to obtain the pollution risk level regions; The cloud computing platform constructs the monitoring resource scheduling model by using a deep reinforcement learning algorithm. The monitoring resource scheduling model takes the pollution risk level regions as the state space, takes the real-time position information and remaining power of the assisted bicycles as constraint conditions, calculates the optimal monitoring path based on the value function iteration method, and the monitoring resource scheduling model assigns inspection routes and sampling frequencies to each assisted bicycle according to the optimal monitoring path, and generates the monitoring task allocation plan including sampling regions and sampling durations.

7. The dynamic traffic environment monitoring method according to claim 1, wherein The cloud computing platform sends the monitoring task allocation plan including the inspection route to the assisted bicycles, calculates the pollution risk index of each road section in the future period according to the pollutant diffusion trend prediction results, and generates a warning message including traffic control strategies, including: The cloud computing platform encodes the monitoring task allocation scheme according to the data transmission protocol, and sends the encoded monitoring task allocation scheme to each of the assisted bicycles based on the wireless communication network. After receiving the monitoring task allocation scheme, the assisted bicycles decode it, plan the sampling trajectory according to the inspection route, and the cloud computing platform tracks the task execution status of each assisted bicycle in real time. When it is detected that the assisted bicycle deviates from the inspection route, it recalculates the optimal monitoring path; The cloud computing platform performs grid division on the road network in the monitoring area based on the pollutant diffusion trend prediction result, calculates the change trend of the pollutant concentration in each grid cell in the future time period, compares the pollutant concentration change trend with the preset pollutant concentration threshold to obtain the pollution risk index, and the cloud computing platform determines the traffic flow regulation area according to the spatial distribution of the pollution risk index, and generates the warning information including the vehicle flow guidance scheme based on the correlation model between traffic flow and pollutant concentration.

8. A dynamic traffic environment monitoring device, characterized in that, The device includes: An environmental acquisition module, which is used to configure an assisted bicycle acquisition network with an environmental monitoring acquisition device. The environmental monitoring acquisition device acquires the particulate matter concentration, nitrogen oxide content, and carbon monoxide concentration, records the sampling position coordinates and sampling timestamps, and the environmental monitoring acquisition device organizes the particulate matter concentration, the nitrogen oxide content, the carbon monoxide concentration, the sampling position coordinates, and the sampling timestamps into a monitoring data packet and uploads it to the cloud computing platform; A pollution prediction module, which is used for the cloud computing platform to perform data preprocessing on the received monitoring data packet, identify outliers for the particulate matter concentration, the nitrogen oxide content, and the carbon monoxide concentration based on the time series clustering algorithm to obtain effective monitoring data. The cloud computing platform constructs a spatio-temporal data matrix according to the sampling position coordinates and the sampling timestamps for the effective monitoring data, extracts features from the spatio-temporal data matrix using a multi-channel convolutional neural network, establishes a pollutant concentration distribution model, and the cloud computing platform trains a pollutant migration model by combining historical monitoring data, inputs the output features of the pollutant concentration distribution model into the pollutant migration model, and generates a pollutant diffusion trend prediction result; A dynamic monitoring module, which is used for the cloud computing platform to divide the pollution risk level area according to the pollutant diffusion trend prediction result, construct a monitoring resource scheduling model based on the deep reinforcement learning algorithm, use the pollution risk level area, the real-time position information and the remaining power of the assisted bicycle as input parameters to generate a monitoring task allocation scheme, and the cloud computing platform sends the monitoring task allocation scheme including the inspection route to the assisted bicycle, calculates the pollution risk index of each road section in the future time period according to the pollutant diffusion trend prediction result, and generates warning information including traffic control strategies.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the dynamic traffic environment monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the dynamic traffic environment monitoring method according to any one of claims 1 to 7.

Images