A smart city regional management method and system based on the Internet of Things

Abstract

The embodiment of the present specification provides a kind of based on the method and system of Internet of Things wisdom city area management.It is to obtain the environmental monitoring data of target area by sensing network platform;Based on the environmental monitoring data, the air pollution situation of the target area is predicted by area prediction model, and the area prediction model is machine learning model;Determine the air pollution situation of the target area in the sub-target area that the target area meets preset condition;Vehicle information is obtained by object platform in the sub-target area;Based on the vehicle information and the air pollution situation of the target area, send the discount information to user platform by service platform, and the discount information is the information related to vehicle driving cost control and / or vehicle driving route discount recommendation, and the discount information is used to the road traffic flow in the sub-target area is introduced.

Description

[0001] Case Analysis

[0002] This application is a divisional application of Chinese application filed on May 23, 2022, with application number 202210559574.7 entitled "A Smart City Air Quality Prediction Method and System Based on the Internet of Things". Technical Field

[0003] This specification relates to the field of Internet of Things (IoT) and regional management, and in particular to a smart city regional management method and system based on IoT. Background Technology

[0004] With social development, air quality has become an increasingly important public concern. Many factors influence air quality, such as vehicle exhaust emissions, factory emissions, dust from construction sites, excavation work, and sandstorms. Especially in cities, air pollution levels can vary significantly between different areas. Predicting future air quality changes within a region is helpful for regulating and coordinating vehicle traffic flow in urban networks. For areas with severe air pollution, identifying potential sources of air pollution is also a crucial task.

[0005] Therefore, a smart city area management method and system based on the Internet of Things is proposed, which can predict air quality information in different areas, identify potential pollution sources, and control and coordinate vehicles in the urban transportation network. Summary of the Invention

[0006] This specification provides one or more embodiments of a smart city area management method based on the Internet of Things (IoT), the method being executed by a management platform. The method includes: acquiring environmental monitoring data of a target area through a sensor network platform; predicting air pollution in the target area based on the environmental monitoring data using a regional prediction model, wherein the regional prediction model is a machine learning model; wherein the target area is a hexagonal area, and the input to the regional prediction model includes temporal characteristics of the target area and other adjacent hexagonal areas; the temporal characteristics include time features, traffic features, and meteorological features; determining sub-target areas within the target area whose air pollution meets preset conditions; acquiring vehicle information entering the sub-target areas through an object platform; and, based on the vehicle information and the air pollution situation in the target area, sending preferential information to a user platform through a service platform, wherein the preferential information is related to vehicle driving cost control and / or vehicle route discount recommendations, and the preferential information is used to divert traffic flow in the sub-target areas.

[0007] This specification provides one or more embodiments of a smart city area management system based on the Internet of Things (IoT). The system includes a user platform, a service platform, a management platform, a sensor network platform, and an object platform. The user platform receives promotional information sent by the service platform. The sensor network platform acquires environmental monitoring data of the target area. The management platform acquires the environmental monitoring data of the target area through the sensor network platform and predicts the air pollution situation of the target area based on the environmental monitoring data using a regional prediction model, where the regional prediction model is a machine learning model. The target area is a hexagonal region, and the input to the regional prediction model includes the temporal characteristics of the target area and other adjacent hexagonal regions. The time-series characteristics include time characteristics, traffic characteristics, and meteorological characteristics; sub-target areas within the target area whose air pollution conditions meet preset conditions are identified; vehicle information entering the sub-target areas is obtained through an object platform; based on the vehicle information and the air pollution conditions of the target areas, preferential information is sent to the user platform through a service platform, the preferential information being information related to vehicle driving cost control and / or vehicle driving route discount recommendations, the preferential information being used to divert road traffic flow in the sub-target areas; the service platform is used to send the preferential information determined based on the vehicle information and the air pollution conditions of the target areas to the user platform; the object platform is used to obtain vehicle information entering the sub-target areas.

[0008] This specification provides one or more embodiments of a smart city area management device based on the Internet of Things, including a processor for executing the above-described method.

[0009] This specification provides one or more embodiments of a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions from the storage medium, the computer performs the above-described method. Attached Figure Description

[0010] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0011] Figure 1 This is a schematic diagram illustrating the application scenarios of the smart city air quality prediction method according to some embodiments of this specification;

[0012] Figure 2 This is a system diagram of a smart city air quality prediction system according to some embodiments of this specification;

[0013] Figure 3This is an exemplary flowchart of a smart city air quality prediction method according to some embodiments of this specification;

[0014] Figure 4 These are exemplary schematic diagrams of regional prediction model structures shown in some embodiments of this specification;

[0015] Figure 5 This is an exemplary flowchart illustrating the identification of suspicious areas based on some embodiments of this specification. Detailed Implementation

[0016] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0017] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way to distinguish different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0018] As indicated in this specification, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0019] Flowcharts are used in this specification to illustrate the operations performed by the system according to embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0020] Figure 1 This is a schematic diagram illustrating the application scenario of the smart city air quality prediction method according to some embodiments of this specification.

[0021] like Figure 1As shown, application scenario 100 of the smart city air quality prediction system may include a processing device 110, a network 120, a database 130, a data acquisition terminal 140, and a user terminal 150. In some embodiments, the components in application scenario 100 may be connected to and / or communicate with each other via network 120 (e.g., wireless connection, wired connection, or a combination thereof). For example, processing device 110 may be connected to database 130 via network 120. As another example, user terminal 150 may be connected to processing device 110 and database 130 via network 120.

[0022] Processing device 110 can be used to process information and / or data related to application scenario 100, such as predicting air pollution conditions, generating prompts or promotional information instructions, etc. In some embodiments, processing device 110 may include one or more processing engines (e.g., a single-chip processing engine or a multi-chip processing engine). By way of example only, processing device 110 may include a central processing unit (CPU). Processing device 110 can process data, information, and / or processing results obtained from other devices or system components, and execute program instructions based on such data, information, and / or processing results to perform one or more functions described in this specification.

[0023] Network 120 can connect the various components of application scenario 100 and / or connect application scenario 100 with external resources. The network enables communication between the components, and with other parts outside application scenario 100, facilitating the exchange of data and / or information. The network can be a local area network (LAN), a wide area network (WAN), the Internet, etc., and can be a combination of various network structures. Database 130 can be used to store data and / or instructions. In some embodiments, database 130 can store data and / or instructions used by processing device 110 to execute or use in order to perform the exemplary methods described in this specification. In some embodiments, database 130 can be connected to network 120 to communicate with one or more components of application scenario 100 (e.g., processing device 110, user terminal 150).

[0024] The data acquisition terminal 140 can be used to collect data and / or information, such as air quality data, meteorological data, and captured images. For example, the data acquisition terminal 140 may include a meteorological satellite 140-1, an air quality detection device 140-2, a land satellite 140-3, etc. In some embodiments, the data acquisition terminal can transmit the collected data and / or information to a processing device via a network.

[0025] User terminal 150 may include one or more terminal devices or software. In some embodiments, user terminal 150 may include mobile phone 150-1, tablet computer 150-2, laptop computer 150-3, etc. In some embodiments, a user can view information and / or input data and / or commands through the user terminal. For example, a user can view information about vehicles that frequently enter a certain area through the user terminal. Another example is that a user can input commands to send text messages through the user terminal, such as sending discount information to drivers, guiding them on alternative routes, or reminding them to avoid areas with severe air pollution.

[0026] It should be noted that application scenario 100 is provided for illustrative purposes only and is not intended to limit the scope of this application. Those skilled in the art can make various modifications or variations based on the description in this specification. For example, application scenario 100 may also include an information source. However, these changes and modifications will not depart from the scope of this application.

[0027] An Internet of Things (IoT) system is an information processing system that includes some or all of the following platforms: a user platform, a service platform, a management platform, a sensor network platform, and an object platform. The user platform is the functional platform for acquiring user-sensed information and generating control information. The service platform connects the management platform and the user platform, providing communication services for both sensing and control information. The management platform coordinates and manages the connections and collaboration between the various functional platforms (such as the user platform and the service platform). The management platform aggregates information from the IoT operating system and provides sensing and control management functions. The service platform connects the management platform and the object platform, providing communication services for both sensing and control information. The user platform is the functional platform for acquiring user-sensed information and generating control information.

[0028] Information processing in an IoT system can be divided into user-perceived information processing and control information processing. Control information can be generated based on user-perceived information. In some embodiments, control information may include user-demand control information, and user-perceived information may include user query information. The processing of perception information involves the object platform acquiring the perception information and transmitting it to the management platform via the sensor network platform. User-demand control information is transmitted from the management platform to the user platform via the service platform, thereby controlling the sending of prompt information.

[0029] In some embodiments, when an IoT system is applied to urban management, it can be referred to as a smart city IoT system.

[0030] Figure 2 This is a system diagram of a smart city air quality prediction system according to some embodiments of this specification. For example... Figure 2 As shown, the smart city air quality prediction system 200 can be implemented based on an Internet of Things (IoT) system. The smart city air quality prediction system 200 includes a user platform 210, a service platform 220, a management platform 230, a sensor network platform 240, and an object platform 250. In some embodiments, the smart city air quality prediction system 200 can be part of or implemented by the processing device 110.

[0031] In some embodiments, the smart city air quality prediction system 200 can be applied to various scenarios for air quality prediction. In some embodiments, the smart city air quality prediction system 200 can acquire meteorological information, traffic information, etc., under various scenarios to obtain air quality prediction values ​​for each scenario. In some embodiments, the smart city air quality prediction system 200 can determine the sources of air pollution based on the acquired air quality prediction values ​​for each scenario.

[0032] Multiple scenarios for air quality prediction can include air quality prediction for a target area, air quality prediction for adjacent areas of the target area, comparing the predicted air quality with the actual air quality in a region, identifying suspicious areas based on the differences, and investigating pollution sources in suspicious areas. It should be noted that the above scenarios are merely examples and do not limit the specific application scenarios of the smart city air quality prediction system 200. Those skilled in the art can apply the smart city air quality prediction system 200 to any other suitable scenario based on the content disclosed in this embodiment.

[0033] In some embodiments, when the smart city air quality prediction system 200 is applied to predict the air quality of a certain area, the management platform 230 acquires the area to be predicted (i.e., the target area), which can be hexagonal. Simultaneously, it acquires six adjacent hexagonal areas of the target area. The management platform 230 acquires relevant data of the target area and the six adjacent hexagonal areas, including meteorological data (such as wind speed, wind direction, air pressure, humidity, temperature, etc.) and traffic data (vehicle flow, congestion, etc.) for each area. The data can be a sequence of historical data from multiple time points up to the current time (e.g., multiple time points at 30-minute or one-hour intervals from the past day or week). The relevant data is processed by a regional prediction model to predict the air quality (i.e., air pollution) of the target area.

[0034] In some embodiments, when the smart city air quality prediction system 200 is applied to predict the air quality of adjacent areas of a target area, it can perform iterative prediction using a regional prediction model. The iterative prediction can be based on the predicted air quality data of each area at a certain time point, predicting the air quality of each area at the next time point. For example, after predicting the air quality of each area at a future first time point, the relevant data of that future first time point will be used as historical time point data for a future second time point, and as input data for the regional prediction model to predict the air quality at that future second time point. For air quality prediction of a target area, introducing the pollution situation of adjacent areas for iterative prediction can make the air quality prediction of the target area more accurate. Furthermore, predictions can be performed sequentially on six adjacent hexagonal areas to obtain the predicted air quality values ​​of the six adjacent hexagonal areas, thereby obtaining the air quality of a wider range of areas, reducing random errors in the data, and improving the accuracy of the target area prediction. For a related description of air quality prediction, see [link to relevant documentation]. Figure 5 The details and related descriptions will not be repeated here.

[0035] In some embodiments, the smart city air quality prediction system 200 is used to compare the predicted air quality of a region with the actual air quality, and to identify suspicious areas based on the difference between the two. The management platform 230 not only acquires the predicted air quality values ​​of the target region and its six adjacent hexagonal regions, but also acquires the actual air quality values ​​of the target region and its six adjacent hexagonal regions. The actual air quality values ​​can be obtained by air quality detection devices 140-2 installed in each region within the target platform 250, and uploaded to the management platform 230 via the sensor network platform 240. The management platform 230 obtains the air quality difference value by comparing the predicted air quality value of each region with the corresponding actual air quality value, and identifies suspicious areas based on the difference value. The suspicious areas may contain additional pollution sources.

[0036] In some embodiments, when the smart city air quality prediction system 200 is applied to investigate pollution sources in suspected areas, the management platform 230 acquires multiple images of the suspected areas. These multiple images may be acquired by Landsat 140-3 in the object platform 250 and uploaded to the management platform 230 via the sensor network platform 240. The management platform 230 can identify the multiple images using a pollution source identification model to determine the category of the pollution source. For example, the pollution source category may be a motor vehicle source (such as motor vehicle exhaust emissions, air pollution including nitrogen dioxide, carbon monoxide, and PM2.5 pollution), an industrial source (such as factory exhaust emissions, air pollution including sulfur dioxide pollution), or a dust source (such as construction site dust, excavation operations, sandstorms, straw burning, etc., pollution including particulate matter pollution), etc.

[0037] In other embodiments, when the smart city air quality prediction system 200 is applied to investigate pollution sources in suspected areas, the management platform 230 can combine other characteristics with an emission determination model to identify suspicious points. These other characteristics can be air quality features, including the concentrations of PM2.5, PM10, ozone, nitrogen dioxide, sulfur dioxide, and carbon monoxide in the air. The suspicious points can refer to the difference between the air quality indicators and the pollution indicators of the pollution source. For example, for the suspected area, if the investigated pollution source is straw burning, it mainly affects the PM2.5 index. If the ozone pollution index in the air quality data of this area is high, it indicates that there are other pollution sources, and further investigation is needed to identify the ozone pollution source.

[0038] Once those skilled in the art understand the principles of this system, they can likely adapt it to any other suitable scenario without departing from those principles.

[0039] The following will use the application of the smart city air quality prediction system 200 in an air quality prediction scenario in a certain area of ​​a city as an example to give a detailed explanation of the smart city air quality prediction system 200.

[0040] User platform 210 can refer to a user-centric platform, including a platform for acquiring user needs and providing information feedback to users. In some embodiments, the user platform is configured to allow users to input commands via a user terminal to query air quality information for a specific area of ​​a city, or to query registration information of vehicles that frequently enter that area. In some embodiments, the user platform is configured to display the queried vehicle registration information and the vehicle owner information via a display terminal. In other embodiments, the user platform is configured as a user terminal to acquire prompts or promotional information input by the user and send it to the target vehicle owner via SMS.

[0041] Service platform 220 can refer to a platform that conveys user needs and control information. It connects user platform 210 and management platform 230. In some embodiments, service platform 220 is configured to receive instructions from users through the user platform to query air quality information and vehicle registration information for a certain area of ​​a city, and to provide the vehicle registration information back to the user. In some embodiments, service platform 220 can receive information from management platform 230, perform data processing operations such as extraction, classification, and reprocessing on the received information to generate valuable information such as statistical data, trend data, and comparative data, and provide corresponding services to users according to their needs. For example, when a user queries air quality information for a certain area, service platform 220, according to the user's request, obtains air quality prediction information for that area from management platform 230, and summarizes the air quality information of adjacent surrounding areas before providing it back to the user. In some embodiments, service platform 220 receives user queries for suspected air pollution sources in a certain area, obtains the investigation results of suspected air pollution sources from management platform 230, and provides them back to the user.

[0042] The management platform 230 can refer to a platform for predicting air quality in different areas of a city. In some embodiments, the management platform 230 can be configured to acquire a target area and adjacent areas, wherein the target area can be a hexagon and the adjacent areas can be six adjacent hexagons. In some embodiments, the management platform 230 can be configured as a regional prediction model to predict the air quality of the target area, with inputs including the hexagonal area and features of the six adjacent hexagonal areas.

[0043] In some embodiments, the feature further includes vehicle pollution data, wherein the vehicle pollution data is acquired by: acquiring road condition prediction data and vehicle registration data of the hexagonal region and other hexagonal regions adjacent to the hexagonal region; processing the road condition prediction data and the vehicle registration data through a vehicle pollution model to predict the vehicle pollution data of the hexagonal region and other hexagonal regions adjacent to the hexagonal region.

[0044] In some embodiments, the management platform 230 may be configured to iteratively predict the air pollution status of the hexagonal region and other adjacent hexagonal regions at a second time point based on the air pollution status of the hexagonal region at a first time point and the air pollution status of other adjacent hexagonal regions through a regional prediction model.

[0045] In some embodiments, the management platform 230 may be configured to acquire deviation data between air pollution conditions and actual air pollution conditions; and based on the deviation data, to identify suspicious areas, wherein the suspicious areas are areas that include pollution sources.

[0046] In some embodiments, the management platform 230 can be configured to send prompt information through the service platform 220 based on the air pollution situation of the target area, including: identifying sub-target areas within the target area whose air pollution situation meets preset conditions; obtaining vehicle information entering the sub-target areas through the object platform 250; and sending prompt information and preferential information through the service platform 220 based on the vehicle information and the air pollution situation of the target area. In some embodiments, by sending information to car owners who frequently enter the target area, prompting them to detour or change their driving routes, further deterioration of the air quality in the area can be avoided. In some embodiments, the management platform can also determine preferential information based on vehicle information and the air pollution situation of the target area, and send the preferential information through the service platform. In some embodiments, the preferential information may be information related to vehicle driving cost control and / or vehicle driving route preferential recommendations. Wherein, vehicle driving cost control may include travel subsidies for using lower-energy-consumption travel modes to increase the cost of using high-energy-consumption travel that aggravates air pollution in the sub-target area. In some embodiments, the preferential information is used to divert road traffic flow in the sub-target area. For example, when a vehicle is traveling within a sub-target area, the discount information could be recommendations for public transportation, low-emission vehicles, or new energy vehicles, along with corresponding cost reductions and subsidies. It could also include coupons (e.g., gas station coupons) or toll exemptions for alternative routes outside the sub-target area. In some embodiments, the discount information can be generated using preset rules. For example, the discount information could be "travel mode + distance traveled in kilometers multiplied by cost savings per unit distance and subsidy value," such as "New energy vehicle: 60 km × (cost savings of 0.35 yuan / km + subsidy of 0.25 yuan / km) = 36 yuan." In some embodiments, after the management platform determines the discount information, it can be sent to the user platform through the service platform. In some embodiments, sending discount information to car owners, such as toll discounts at other toll stations or gas stations, attracts car owners, diverts traffic flow, and alleviates air pollution caused by vehicle exhaust emissions in the target area.

[0047] The sensor network platform 240 can refer to a functional platform that manages communications for urban air quality forecasting. In some embodiments, the sensor network platform 240 can connect to the management platform 230 and the object platform 250 to achieve communication functionality. In some embodiments, the sensor network platform may include multiple sensor network sub-platforms. Different sensor network sub-platforms provide corresponding communication channels for different sensing devices. For example, the meteorological satellite sensing sub-platform communicates with meteorological satellites to acquire meteorological data. In some embodiments, different sensor network sub-platforms are used according to the type of different acquisition devices, making the sensor network platform 240 more targeted in acquiring and processing the collected information. At the same time, multiple communication channels alleviate the pressure of data uploading to the sensor network platform and improve the efficiency of system operation.

[0048] The object platform 250 can refer to a platform that acquires external sensory information through various sensing devices. For example, the object platform 250 includes an air quality monitoring device installed in a certain area of ​​a city to acquire air quality information for that area and upload it to the management platform 230 via a sensor network platform 240 for comparison processing of predicted air quality values. Another example is a meteorological satellite, used to acquire meteorological data for a certain area and upload the meteorological data to the management platform 230 via the sensor network platform 240 for predicting regional air quality. Yet another example is a Landsat satellite, used to acquire images of a certain area and upload the images to the management platform 230 via the sensor network platform 240 for investigating and processing suspected air pollution sources.

[0049] It should be noted that the above description of the system and its components is for convenience only and should not be construed as limiting this specification to the embodiments described. It is understood that those skilled in the art, after understanding the principles of the system, may arbitrarily combine the various components or construct subsystems connected to other components without departing from these principles. For example, the components may share a single database, or each component may have its own separate database. Such modifications are all within the scope of this specification.

[0050] Figure 3 This is an exemplary flowchart illustrating a smart city air quality prediction method according to some embodiments of this specification. Figure 3 As shown, process 300 includes the following steps. In some embodiments, process 300 may be executed by a management platform.

[0051] S310, acquire environmental monitoring data of the target area, including at least one of air quality data, weather data, and satellite image data.

[0052] The target area can be a specific region where air quality forecasting is required. For example, the target area can be a city, such as Shanghai, or a specific district within a city, such as Jing'an District. In some embodiments, the management platform can divide an area into multiple honeycomb-shaped hexagonal regions, and the target area can be one or more of these hexagonal regions. In some embodiments, the target area can be adjacent to other areas; for example, when a hexagonal region is used as the target area, it can be adjacent to six other hexagonal regions. In some embodiments, the target area can also be other shapes. For example, the target area can also be rectangular, and it can be adjacent to eight other rectangular regions.

[0053] Environmental monitoring data can be any monitored data related to the environment. For example, environmental monitoring data can include at least one of the following: air quality data, weather data, and satellite imagery data.

[0054] Air quality data can reflect the degree of air pollution. For example, air quality data can include the air pollution index, the concentration of various pollutants in the air (such as nitrogen oxides, carbon monoxide, and inhalable particulate matter).

[0055] Weather data can reflect the climate and meteorological conditions within a certain time and area. For example, weather data can include temperature, humidity, precipitation, wind speed, etc.

[0056] Satellite imagery data can be data fed back by satellites that conduct meteorological observations of Earth. For example, satellite imagery data can include image data such as satellite cloud images.

[0057] In some embodiments, the aforementioned environmental monitoring data can be acquired through different platforms. For example, satellite imagery data can be acquired through a satellite platform; air quality data can be acquired through an air quality monitoring platform; and weather data can be determined through meteorological websites or sensors related to weather index measurements.

[0058] S320, based on environmental monitoring data, predicts air pollution in target areas using a regional prediction model, which is a machine learning model.

[0059] Regional prediction models can be models used to predict air pollution conditions in a target area. For example, regional prediction models can be deep neural network (DNN) models, recurrent neural network (RNN) models, or combinations thereof.

[0060] Air pollution information can be any information related to air pollution in the target area. For example, air pollution information can include pollution type such as industrial pollution, radioactive pollution, biological pollution, etc.; pollution source such as chemical plant exhaust, straw burning fumes, etc.; and pollution level such as urban air quality level, and the content of harmful substances in the air such as PM2.5 levels. In some embodiments, air pollution information includes the actual air quality obtained through air quality sensors.

[0061] In some embodiments, the management platform can predict air pollution levels in a target area based on environmental monitoring data in various ways. For example, the air pollution level in the target area can be predicted based on the mapping relationship between environmental monitoring data and the air pollution level in the target area.

[0062] In some embodiments, the air pollution situation in the target area can be predicted based on a regional prediction model, wherein the regional prediction model can be a machine learning model.

[0063] For more information on regional prediction models for predicting air pollution in target areas, please refer to [link / reference]. Figure 4 Some specific details.

[0064] S330 determines the alert information based on the air pollution situation in the target area, and sends the alert information to the user platform through the service platform.

[0065] The notification message can be information related to the air pollution situation in the target area. For example, the notification message can be any form of information reflecting the pollution source, pollution type, and pollution severity level of the target area, such as text, images, or voice. In some embodiments, the notification message can be presented and sent to the user on the user's terminal. In some embodiments, the notification message may also include images or videos related to the pollution source, the pollution source category, and the corresponding pollution treatment method.

[0066] In some embodiments, the prompt message can be generated using preset rules. For example, the prompt message can be a preset text rule of "pollution type + pollution source + pollution severity level," such as "pollution type: biological pollution; pollution source: farmland straw burning; pollution severity level: severe." In some embodiments, the prompt message can also include other preset rule forms, such as preset rules related to images or preset rules related to voice. In some embodiments, the prompt message can be sent from the service platform to the user platform via a network. For example, the prompt message can be an SMS notification sent from the service platform to the user platform, thereby allowing the vehicle owner to receive it through their terminal device. In some embodiments, promotional information can also be sent through the service platform based on vehicle information and the air pollution situation in the target area. For details on promotional information, please refer to [link to relevant information]. Figure 2The corresponding section's content. In some embodiments, the discount amount or the intensity of the reminder information in the discount information can be determined based on vehicle information and the air pollution situation in the target area. In some embodiments, it can be determined based on the air quality pollution level of the area, such as the air pollution index (air pollution index 101-150 is light pollution; 151-200 is moderate pollution; 201-300 is heavy pollution; and above 300 is severe pollution). The more severe the pollution level, the greater the discount amount and the stronger the reminder.

[0067] The smart city air quality prediction method based on the Internet of Things described in some embodiments of this specification can realize air quality prediction based on machine learning models. The machine learning models can process environmental monitoring data obtained from various platforms and channels, improve the comprehensiveness of prediction and air quality-related information acquisition, and improve the accuracy of air quality prediction results.

[0068] It should be noted that the above description of process 300 is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to process 300 under the guidance of this specification. However, these modifications and changes remain within the scope of this specification. For example, process 300 may also include a pretreatment step.

[0069] Figure 4 This is an exemplary schematic diagram of a regional prediction model structure according to some embodiments of this specification. Structure 400 as shown... Figure 4 As shown. In some embodiments, structure 400 can be implemented through a management platform.

[0070] In some embodiments, the regional prediction model may include at least one regional model and an air quality assessment model, wherein each regional model corresponds to one region. For example, at least one regional model may be seven regional models, namely a first regional model, a second regional model, ... a seventh regional model, each model corresponding to one region, wherein the seventh regional model may correspond to the target region, and the first regional model, the second regional model, ... the sixth regional model may correspond to other regions adjacent to the target region (i.e., corresponding to the first region, the second region, ... the seventh region, respectively).

[0071] It is understood that the descriptions of the order and number of each region model in this specification are for illustrative purposes only and do not imply any limitation on the order, number, function, or corresponding region of the region models. For example, the first region model could be the seventh region model mentioned above, and the second region model could correspond to the target region, etc. Furthermore, the number of region models can be determined based on the shape of the target region; for example, if the shape of the region is rectangular, then the number of region models is 5.

[0072] The input to the regional prediction model can include features of the target region (hexagonal region) and features of other hexagonal regions adjacent to the target region (hexagonal region), and the output of the regional prediction model can include the air quality of the target region.

[0073] In some embodiments, the features of the target region (hexagonal region) can be the temporal features of the target region, and the features of other hexagonal regions adjacent to the target region can be the temporal features of other hexagonal regions. Temporal features can be the traffic and meteorological features of a region at each point in time over a period of time; that is, temporal features can include time features, traffic features, and meteorological features. For example, a time step of 2 hours can be set within a day, and a time point is obtained every 2 hours. Within the time period of a day, traffic features such as traffic flow and congestion, meteorological features such as wind speed, wind direction, air pressure, humidity, and temperature at each time point, and the corresponding time point (i.e., time features) are used as input to the model. In some embodiments, temporal features can include first temporal features, second temporal features, ..., seventh temporal features, etc., used to represent the corresponding temporal features within different regions.

[0074] The first region model can be a model used to extract features from the first region. For example, the first region model can be a deep neural network model.

[0075] The input to the first region model may include temporal features, and the output may include the first temporal feature vector.

[0076] A temporal feature vector can be a vector that reflects changes in traffic and meteorological characteristics of a region over a certain period of time. For example, the elements of the vector can include a time point, as well as characteristics such as traffic flow, congestion, wind speed, wind direction, air pressure, humidity, and temperature at that time point.

[0077] In some embodiments, the regional prediction model may further include a second regional model, a third regional model, ... a seventh regional model, etc., each corresponding to a different region. The model type, input, and output of these regional models may be similar to those of the first regional model. For example, the second regional model may be a deep neural network model, and its input may include temporal features, while its output may include a second temporal feature vector, and so on. In summary, the temporal feature vectors corresponding to the seven regions can be obtained through the above seven regional models. Each temporal feature vector can reflect various features of the corresponding region, and the seventh temporal feature vector can correspond to the target region.

[0078] Air quality assessment models can be used to confirm air pollution levels. For example, an air quality assessment model can be a deep neural network model.

[0079] The input to the air quality assessment model can include the first time-series feature vector, the second time-series feature vector, ... the seventh time-series feature vector, and the output can include the air pollution status of the target area.

[0080] In some embodiments, the regional prediction model can be obtained through joint training. For example, training sample data, which represents the temporal features of the corresponding regions, is input into a first regional model, a second regional model, ... a seventh regional model, to obtain output first temporal feature vectors, second temporal feature vectors, ... seventh temporal feature vectors. These temporal feature vectors are then used as training sample data for an air quality assessment model, input into the model to obtain the air pollution situation in the target region, and the output of the air quality assessment model is validated using the air pollution situation in the target region. For example, the actual air quality obtained by air quality monitoring devices installed in the target region can be used as a label for the air quality assessment model for validation.

[0081] For example, the training sample data includes the temporal features of the corresponding region and the temporal feature vector. The training sample data is input into the first region model, the second region model, ..., the seventh region model, and the sample temporal feature vector is input into the air quality judgment model. The outputs of the first region model, the second region model, ..., the seventh region model serve as the input to the air quality judgment model, and the label represents the air pollution status of the target region. In some embodiments, the training label can be the actual air quality obtained through an air quality sensor. During training, a loss function is established based on the air pollution status of the target region and the outputs of the first region model, the second region model, ..., the seventh region model to update the model parameters.

[0082] In some embodiments, the training sample data may include at least temporal features of different regions. Labels may be the air pollution status of the target region. Labels may be obtained through manual annotation.

[0083] In some embodiments, the input to the regional prediction model may further include vehicle pollution data (also referred to as vehicle pollution features). In some embodiments, the acquisition of vehicle pollution data includes: acquiring road condition prediction data and vehicle registration data for a hexagonal region and other hexagonal regions adjacent to the hexagonal region; processing the road condition prediction data and vehicle registration data through a vehicle pollution model to predict vehicle pollution data for the hexagonal region and other hexagonal regions adjacent to the hexagonal region.

[0084] Vehicle pollution data can be data on air pollution caused by vehicles within a certain time period. For example, vehicle pollution data may include data such as vehicle exhaust emission levels, exhaust emission components, and exhaust emission distribution. In some embodiments, vehicle pollution data can be determined using a vehicle pollution model. In some embodiments, some or all of the vehicle pollution data may be time-series features.

[0085] A vehicle pollution model can be a model that determines vehicle pollution data. For example, a vehicle pollution model can be a deep neural network model, etc. In some embodiments, a vehicle pollution model may include a first vehicle pollution model, a second vehicle pollution model, ... a seventh vehicle pollution model, etc., used to represent models corresponding to vehicle pollution data in different regions, all of which belong to vehicle pollution models with the same parameters.

[0086] The inputs to a vehicle pollution model may include road condition prediction data and vehicle registration data, and the output of the vehicle pollution model may include vehicle pollution data. In some embodiments, the output vehicle pollution data may be represented by air quality data; therefore, vehicle pollution data can also be understood as air quality data under the conditions corresponding to the input road condition prediction data and vehicle registration data.

[0087] Traffic forecasting data can be statistical data on traffic conditions within a region over a historical period. For example, traffic forecasting data can include the number of vehicles on the road, the degree of road congestion, etc. Road traffic flow data includes characteristics such as exhaust emissions.

[0088] Vehicle registration data can be information related to vehicles within a specific historical period. For example, vehicle registration data may include the vehicle's power type (fuel, electric, etc.), emissions, fuel consumption, etc.

[0089] In some embodiments, the vehicle pollution model can be a DNN, whose input can be road traffic flow data and whose output can be air quality data (used to represent vehicle pollution data). The road traffic flow data includes features such as road traffic volume, the proportion of fuel-powered vehicles, and exhaust emissions.

[0090] In some embodiments, the vehicle pollution model can be obtained through joint training with a regional prediction model. For example, training sample data, namely historical road condition prediction data and historical vehicle registration data, is input into the vehicle pollution model to obtain the output historical vehicle pollution data; then, the aforementioned historical vehicle pollution data is used as training sample data for the regional prediction model and input into the regional prediction model to obtain the air pollution situation of the target area; the output of the regional prediction model is validated using the air pollution situation of the target area; and validation data of the vehicle pollution data output by the vehicle pollution model is obtained by utilizing the backpropagation characteristics of the neural network model, and the validation data is used as labels to train the vehicle pollution model.

[0091] For example, the training sample data includes historical road condition prediction data, historical vehicle registration data, and historical vehicle pollution data. The training sample data is input into the vehicle pollution model, and the historical vehicle pollution data is input into the regional prediction model. The output of the vehicle pollution model is used as the input of the regional prediction model, and the label is the air pollution situation of the target area. During the training process, a loss function is established based on the air pollution situation of the target area and the output of the vehicle pollution model to update the parameters of the model.

[0092] In some embodiments, the training sample data may include at least historical traffic prediction data and historical vehicle registration data. Tags may represent air pollution levels in the target area. Tags may be obtained through manual annotation or determined based on monitoring by sensors on the relevant roads, such as air quality sensors.

[0093] In some embodiments, the prediction method of the regional prediction model can be iterative prediction, which includes: based on the air pollution situation of the hexagonal region and other hexagonal regions adjacent to the hexagonal region at a first time point, iteratively predicting the air pollution situation of the hexagonal region and other hexagonal regions adjacent to the hexagonal region at a second time point through the regional prediction model.

[0094] The first time point can be the point in time when a preliminary air pollution forecast is made within a certain period. The second time point can be the point in time when a second air pollution forecast is made within a certain period. For example, the first time point can be expressed as a specific moment such as 9:00, or a specific day, and the second time point can be expressed similarly.

[0095] In some embodiments, iterative prediction may involve obtaining the air pollution situation and time-series characteristics corresponding to the first time point from the prediction process at the first time point, and using the air pollution situation and time-series characteristics corresponding to the first time point as input to the prediction process at the second time point. In some embodiments, the prediction process may also include a third time point, a fourth time point, etc., and the prediction processes for the above time points are the same as those for the prediction processes at the first and second time points.

[0096] In some embodiments, the iteration may terminate when the number of time points reaches a preset number. For example, if the preset number of time points is 5, the iterative prediction ends when the prediction at the fifth time point is completed.

[0097] The regional prediction model described in some embodiments of this specification can be used to predict future air pollution conditions. By incorporating statistical data of vehicles within the region, the impact of vehicle exhaust on air quality can be considered. In addition, the model is obtained through joint training, which can reduce the number of training samples, simplify the training process, and avoid the problem of difficulty in obtaining labels for each regional model.

[0098] Figure 5 This is an exemplary flowchart illustrating the identification of suspicious areas based on some embodiments of this specification. Figure 5 As shown, process 500 includes the following steps. In some embodiments, process 500 may be executed by a management platform.

[0099] S510, Obtain deviation data between air pollution conditions and actual air pollution conditions. In some embodiments, the actual air pollution conditions of the target area are obtained through a sensor network platform, and the deviation data between the air pollution conditions predicted by the regional prediction model and the actual air pollution conditions are determined.

[0100] Deviation data can reflect differences between predicted and actual air pollution levels. For example, deviation data may include a significant discrepancy between the predicted PM2.5 concentration and the actual PM2.5 concentration during a given time period, exceeding a deviation threshold. Deviation data may also include a predicted pollution type of biological pollution during a given time period, while the actual air pollution during that time period is a combination of biological and industrial pollution.

[0101] In some embodiments, deviation data can be obtained by comparing predicted air pollution with actual air pollution through a management platform, wherein the actual air pollution can be obtained through an air quality detection device or by manual input.

[0102] S520 identifies suspicious areas based on deviation data.

[0103] A suspected area can be an area where other sources of pollution may exist. For example, a suspected area can be one or more of the hexagonal areas mentioned above, or it can be a specific address, such as a factory.

[0104] In some embodiments, suspicious areas can be identified using environmental monitoring data. For example, areas where the haze concentration reaches a haze concentration threshold can be identified as suspicious areas using satellite imagery data, or areas where the concentration of dust in the air reaches a dust concentration threshold can be identified as suspicious areas using sensors. In some embodiments, suspicious areas can be identified using vehicle-related data such as vehicle pollution data and road condition prediction data. For example, areas with high vehicle density can be identified as suspicious areas based on the road network distribution of vehicles. In some embodiments, suspicious areas can also be identified through manual judgment.

[0105] S530 acquires image data of suspicious areas and identifies the source of pollution through image recognition.

[0106] In some embodiments, the management platform can extract image features from image data and combine them with other features to determine the pollution source using an emission determination model. In some embodiments, the emission determination model can be a component layer of a pollution identification model, such as a judgment layer. In some embodiments, the management platform acquires image data of suspicious areas and processes the image data of suspicious areas using a pollution source identification model to determine the pollution source. The pollution source identification model may include an image feature extraction layer and a judgment layer. The image feature extraction layer is used to extract image feature vectors of suspicious areas based on the image data of the suspicious areas, and the judgment layer is used to determine the pollution source based on the image feature vectors of the suspicious areas and air quality characteristics.

[0107] A pollution source identification model can be a model used to determine the category of a pollution source. For example, a pollution source identification model can be a convolutional neural network model, a deep neural network model, or a combination thereof.

[0108] In some embodiments, the pollution source identification model may include an image feature extraction layer and a judgment layer.

[0109] The image feature extraction layer can be a model used to extract features in an image that are related to pollution sources. For example, the image feature extraction layer can be a convolutional neural network model.

[0110] The input to the image feature extraction layer can include image data of the suspected region, and the output can include an image feature vector. The image feature vector can be a vector reflecting the features contained in the image. For example, elements of the image feature vector can include exhaust emission characteristics of pollution sources, building features, scale characteristics, etc.

[0111] The decision layer can be a model used to ultimately determine the pollution source category. For example, the decision layer can be a deep neural network model.

[0112] The input to the decision layer can include image feature vectors and air quality data, and the output can include pollution source categories. Pollution source categories can include vehicle sources (such as vehicle exhaust emissions, including nitrogen dioxide, carbon monoxide, and PM2.5 pollution), dust sources (such as construction site dust, excavation operations, sandstorms, and straw burning, including PM2.5 and PM10 pollution), and industrial sources (such as factory exhaust emissions, including sulfur dioxide pollution).

[0113] In some embodiments, the pollution source identification model can be obtained through joint training of an image feature extraction layer and a judgment layer. For example, training sample data, i.e., historical image data, is input into the image feature extraction layer to obtain the output historical image feature vector; then, the aforementioned historical image feature vector, along with historical air quality data, is used as training sample data for the judgment layer and input into the judgment layer to obtain the pollution source category. The sample pollution source category is used to verify the output of the judgment layer; utilizing the backpropagation characteristics of the neural network model, verification data of the historical image feature vector output by the image feature extraction layer is obtained, and this verification data is used as a label to train the image feature extraction layer.

[0114] For example, the training sample data includes historical image data, historical image feature vectors, and historical air quality data. The training sample data is input into the image feature extraction layer, and the historical image feature vectors and historical air quality data are input into the judgment layer. The output of the image feature extraction layer is used as the input of the judgment layer, and the label is the pollution source category. During the training process, a loss function is established based on the pollution source category and the output of the image feature extraction layer to update the parameters of the model.

[0115] In some embodiments, the training sample data may include at least historical image data and historical air quality data. Labels may be pollution source categories. Labels may be obtained through manual annotation or determined based on actual pollution source categories.

[0116] In some embodiments of this specification, suspicious areas are identified based on deviation data between air pollution conditions and actual air pollution conditions. The pollution source is determined by image recognition of the suspicious area's image data, and the pollution source is determined by combining other features with an emission determination model. This approach helps to accurately, objectively, intelligently, and efficiently identify areas with additional pollution sources, and helps to avoid the threat of potential pollution sources.

[0117] It should be noted that the above description of process 500 is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to process 500 under the guidance of this specification. However, these modifications and changes remain within the scope of this specification. For example, process 500 may also include a pretreatment step.

[0118] Some embodiments of this specification also disclose a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer runs the above-described smart city air quality prediction method based on the Internet of Things.

[0119] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0120] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Moreover, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0121] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the spirit and scope of the embodiments described herein. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely using software solutions, such as installing the described system on existing servers or mobile devices.

[0122] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0123] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters should take into account specified significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of range in some embodiments of this specification are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0124] For each patent, patent application, patent application publication, and other material, such as articles, books, specifications, publications, and documents, referenced in this specification, the entire contents of which are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this specification, as well as documents that limit the broadest scope of the claims in this specification (currently or subsequently appended to this specification). It should be noted that in the event of any inconsistency or conflict between the descriptions, definitions, and / or terminology used in the supplementary materials to this specification and the content of this specification, the descriptions, definitions, and / or terminology used in this specification shall prevail.

[0125] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A smart city area management method based on Internet of Things, characterized in that, The method is executed by the management platform, and the method includes: Environmental monitoring data for the target area is acquired through a sensor network platform; Based on the environmental monitoring data, the air pollution situation in the target area is predicted using a regional prediction model, which is a machine learning model. The target area is a hexagonal region, and the input to the regional prediction model includes the temporal characteristics of the target area and other adjacent hexagonal regions. The temporal characteristics include time features, traffic features, and meteorological features. The input to the regional prediction model also includes vehicle pollution data, which includes vehicle exhaust emission levels, vehicle exhaust emission components, and vehicle exhaust emission distribution. The vehicle pollution data is acquired by: acquiring road condition prediction data and vehicle registration data for the target area and other adjacent hexagonal areas; processing the road condition prediction data and vehicle registration data using the vehicle pollution model to predict vehicle pollution data for the target area and other adjacent hexagonal areas; wherein the vehicle pollution model is a deep neural network model, the input to the vehicle pollution model includes the road condition prediction data and vehicle registration data, and the output of the vehicle pollution model includes the vehicle pollution data, which is represented by air quality data and is part of the road condition prediction data. The air quality data corresponding to the vehicle registration data is obtained by the vehicle pollution model through joint training with the regional prediction model. The joint training includes: using historical traffic prediction data, historical vehicle registration data, and historical vehicle pollution data as training sample data; inputting the historical traffic prediction data and historical vehicle registration data from the training sample data into the vehicle pollution model; inputting the historical vehicle pollution data from the training sample data into the regional prediction model to obtain the air pollution situation of the target area; validating the output of the regional prediction model using the air pollution situation of the target area; and using the backpropagation characteristics of the deep neural network model to obtain validation data of the vehicle pollution data output by the vehicle pollution model, using the validation data as labels to train the vehicle pollution model. The training sample data is input into the vehicle pollution model, and the historical vehicle pollution data is input into the regional prediction model. The output of the vehicle pollution model is used as the input of the regional prediction model. The label is the air pollution situation of the target area. During the training process, a loss function is established based on the air pollution situation of the target area and the output of the vehicle pollution model to update the model parameters. Identify sub-target areas within the target area whose air pollution conditions meet preset conditions; The vehicle information entering the sub-target area is obtained through the object platform; Based on the vehicle information and the air pollution situation in the target area, the service platform sends preferential information to the user platform. The preferential information is related to vehicle driving cost control and / or preferential recommendations for vehicle driving routes. The preferential information is used to divert road traffic flow in the sub-target area.

2. The method of claim 1, wherein, The method further includes: The actual air pollution situation in the target area is obtained through the sensor network platform, and the deviation data between the air pollution situation predicted by the regional prediction model and the actual air pollution situation is determined. Based on the deviation data, suspicious areas are identified.

3. The method of claim 2, wherein, The method further includes: Acquire image data of the suspected area, and process the image data of the suspected area using a pollution source identification model to determine the pollution source; Based on the air pollution situation in the target area and / or the pollution sources in the suspected area, a prompt message is generated and sent to the user platform through the service platform.

4. An Internet of Things-based smart city area management system, characterized by, The system includes a user platform, a service platform, a management platform, a sensor network platform, and an object platform. The user platform is used to receive promotional information sent by the service platform; The sensor network platform is used to acquire environmental monitoring data of the target area; The management platform is used for: Environmental monitoring data for the target area is acquired through a sensor network platform; Based on the environmental monitoring data, the air pollution situation in the target area is predicted using a regional prediction model, which is a machine learning model. The target area is a hexagonal region, and the input to the regional prediction model includes the temporal characteristics of the target area and other adjacent hexagonal regions. The temporal characteristics include time features, traffic features, and meteorological features. The input to the regional prediction model also includes vehicle pollution data, which includes vehicle exhaust emission levels, vehicle exhaust emission components, and vehicle exhaust emission distribution. The vehicle pollution data is acquired by: acquiring road condition prediction data and vehicle registration data for the target area and other adjacent hexagonal areas; processing the road condition prediction data and vehicle registration data using the vehicle pollution model to predict vehicle pollution data for the target area and other adjacent hexagonal areas; wherein the vehicle pollution model is a deep neural network model, the input to the vehicle pollution model includes the road condition prediction data and vehicle registration data, and the output of the vehicle pollution model includes the vehicle pollution data, which is represented by air quality data and is part of the road condition prediction data. The air quality data corresponding to the vehicle registration data is obtained by the vehicle pollution model through joint training with the regional prediction model. The joint training includes: using historical traffic prediction data, historical vehicle registration data, and historical vehicle pollution data as training sample data; inputting the historical traffic prediction data and historical vehicle registration data from the training sample data into the vehicle pollution model; inputting the historical vehicle pollution data from the training sample data into the regional prediction model to obtain the air pollution situation of the target area; validating the output of the regional prediction model using the air pollution situation of the target area; and using the backpropagation characteristics of the deep neural network model to obtain validation data of the vehicle pollution data output by the vehicle pollution model, using the validation data as labels to train the vehicle pollution model. The training sample data is input into the vehicle pollution model, and the historical vehicle pollution data is input into the regional prediction model. The output of the vehicle pollution model is used as the input of the regional prediction model. The label is the air pollution situation of the target area. During the training process, a loss function is established based on the air pollution situation of the target area and the output of the vehicle pollution model to update the model parameters. Identify sub-target areas within the target area whose air pollution conditions meet preset conditions; The vehicle information entering the sub-target area is obtained through the object platform; Based on the vehicle information and the air pollution situation in the target area, the service platform sends preferential information to the user platform. The preferential information is related to vehicle driving cost control and / or preferential recommendations for vehicle driving routes. The preferential information is used to divert traffic flow in the sub-target area. The service platform is used to send the preferential information determined based on the vehicle information and the air pollution situation in the target area to the user platform; The object platform is used to obtain vehicle information that has entered the sub-target area.

5. The system according to claim 4, characterized in that, The management platform is also used for: The actual air pollution situation in the target area is obtained through the sensor network platform, and the deviation data between the air pollution situation predicted by the regional prediction model and the actual air pollution situation is determined. Based on the deviation data, suspicious areas are identified.

6. The system according to claim 5, characterized in that, The management platform is also used for: Acquire image data of the suspected area, and process the image data of the suspected area using a pollution source identification model to determine the pollution source; Based on the air pollution situation in the target area and / or the pollution sources in the suspected area, a prompt message is generated and sent to the user platform through the service platform.

7. A smart city area management device based on the Internet of Things, comprising a processor for executing the method of any one of claims 1 to 3.

8. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer performs the method as described in any one of claims 1 to 3.

Images