Digital twin smart city management system based on AI model

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a digital twin smart city management system based on an AI model, and the system comprises the steps: collecting the city data of a target city; constructing a digital twinborn body of the target city by using the collected city data and a modeling technology; deploying a plurality of AI models on the digital twin; the AI model comprises a machine learning model and a deep learning model, and the machine learning model is used for processing and analyzing the acquired city data and determining the real-time running state and development demand of a city corresponding to the city data; the deep learning model is used for processing complex city data and optimizing city management decisions; through interaction between the visual interface and the digital twin smart city management system, decision suggestion information obtained by analysis of each AI model in a plurality of AI models is obtained, digital modeling of a physical entity of a city is realized, the running state of the city can be monitored in real time, and the management efficiency and service level of the city are improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular, to a digital twin smart city management system based on an AI model. Background Art

[0002] With the continuous development of artificial intelligence technology and the continuous improvement of urban management requirements, the digital twin smart city management system will play an increasingly important role in the fields of urban planning, traffic management, environmental protection, public safety, etc. In the future, the digital twin smart city management system will become an important means of urban management and provide strong support for the sustainable development of cities.

[0003] With the progress of society and the development of science and technology, the urbanization process has been accelerating, the scale of cities has been increasing, and new challenges have been posed to urban management.

[0004] In recent years, the concept of smart cities has been put forward, aiming to improve the service level of cities. However, there is no complete solution to how to effectively manage cities to improve the service level. Summary of the Invention

[0005] The main purpose of this application is to provide a digital twin smart city management system based on an AI model, which is used to solve the technical problem of how to effectively manage cities to improve the service level.

[0006] To achieve the above object, the present invention provides a digital twin smart city management system based on an AI model, including: collecting urban data of a target city, where the urban data includes real-time data collected by terminal devices installed in the city, static data and dynamic data obtained from a business system that establishes a communication connection with the digital twin smart city management system, and citizen data collected from the network; Constructing a digital twin of the target city by using the collected urban data and modeling technology; Deploying multiple AI models on the digital twin, where the multiple AI models are pre-trained through the urban data; the AI models include machine learning models and deep learning models, the machine learning models are used to process and analyze the obtained urban data to determine the real-time operation status and development needs of the city corresponding to the urban data; the deep learning models are used to process complex urban data and optimize urban management decisions; Interacting with the digital twin smart city management system through a visualization interface to obtain decision-making advice information analyzed by each of the multiple AI models.

[0007] Optionally, constructing the digital twin of the target city by using the collected city data and modeling techniques includes: Performing data cleaning on the city data to obtain cleaned data; Classifying the cleaned data according to different types and uses of the city data to obtain at least one data set; Selecting corresponding data standardization processing methods for each data set according to the category of each data set in the at least one data set; Performing data standardization processing on each data set according to the corresponding data standardization processing methods to obtain at least one standardized data set; Selecting corresponding modeling techniques for each standardized data set according to the category of each standardized data set; Constructing the digital twin of the target city according to each standardized data set and the corresponding modeling technique in the at least one standardized data set.

[0008] Optionally, constructing the digital twin of the target city according to each standardized data set and the corresponding modeling technique in the at least one standardized data set includes: Based on the standardized data set of infrastructure in the at least one standardized data set, constructing a digital twin of the physical space layout of the target city by using 3D modeling techniques; Constructing the operation logic of the target city by using system dynamics modeling techniques and / or Agent-Based Modeling techniques according to the at least one standardized data set; Based on the digital twin and the operation logic, constructing a digital twin model of the target city and using the digital twin model as the digital twin of the target city.

[0009] Optionally, deploying multiple AI models on the digital twin includes: Determining the correspondence between the category of each standardized data set and each AI model in the multiple AI models; For different AI models deployed on the digital twin, determining the interface information for each AI model to receive data, where the interface information includes the format, type, and transmission frequency of the data transmitted by the corresponding interface; Selecting a corresponding transmission protocol according to the data reception requirement information of each AI model, the corresponding interface information, and the deployment environment of the digital twin; Setting up a data transmission channel from the data collection end to each AI model of the digital twin according to the selected transmission protocol; When data arrives at the corresponding input interface of the digital twin via a transmission channel, the data is parsed according to pre-determined interface information; Integrate the parsed valid data into the corresponding AI model in the digital twin.

[0010] Optionally, deploying multiple AI models on the digital twin includes: In response to determining that the data format and features provided by the digital twin match the AI model to be deployed, further integrate the urban data collected by the digital twin to obtain the input data for the AI model; Select a target AI model according to the application scenario information and expected function information of the digital twin; Pre-evaluate the target AI model to determine the basic model information of the target AI model, where the basic model information includes accuracy, robustness, and computing resource requirements; Determine the overall requirement information of the computing resources for deploying multiple AI models on the digital twin according to the basic model information of each AI model in the multiple AI models; Determine the device information of the target hardware device according to the overall requirement information; Deploy the multiple AI models on the digital twin according to a pre-set model integration framework and the model parameters determined for each AI model in the multiple AI models during the model training process.

[0011] Optionally, the system further includes an urban planning module; The urban planning module includes a visualization sub-module and a construction management sub-module; The visualization sub-module is used to create a three-dimensional model of the city in a virtual environment using digital twin technology and the urban data; The construction management sub-module is used to monitor the construction progress in real time, compare the difference between the planned progress and the actual progress, and adjust the construction arrangement.

[0012] Optionally, the system further includes an energy management module; The energy management module includes an energy detection and analysis sub-module and an energy optimization scheduling sub-module; The energy detection and analysis sub-module is used to monitor the urban energy system in real time, transmit the detected energy data to the visualization sub-module for visual display, where the urban energy system includes a power system, a gas system, and a water supply system; analyze the detected energy data to determine the cause information of energy waste; and transmit the cause information to the energy optimization scheduling sub-module; The energy optimization scheduling sub-module is used to optimize the energy scheduling information according to the obtained energy demand information and energy supply situation.

[0013] Optionally, the system further includes a public safety management module; The public safety management module includes a safety monitoring and early warning sub-module and an emergency rescue command sub-module; The safety monitoring and early warning sub-module is used to comprehensively monitor the urban public safety and obtain safety information in real time; based on the safety information, generate event-related information and early warning information, and transmit the event-related information to the emergency rescue command sub-module, where the event-related information includes the location information of the event, the influence range information of the event, and the severity; The emergency rescue command sub-module is used to generate emergency plan information according to the event-related information; optimize the allocation and scheduling of rescue resources according to the emergency plan information.

[0014] Optionally, the urban data includes traffic flow data, and the system further includes: Real-time monitor the traffic flow data of urban roads; According to the traffic flow data, analyze the causes and locations of traffic congestion, and transmit the traffic flow data to the visualization sub-module for visualization display; According to the traffic flow data, adjust the working time of traffic lights and optimize the traffic signal timing plan; According to the optimized traffic signal timing plan, control the working time of traffic lights in the city.

[0015] Optionally, the urban data includes urban environmental detection data, and the system further includes: Real-time monitor the urban environment, obtain urban environmental detection data, and transmit the urban environmental detection data to the visualization sub-module for visualization display, where the urban environmental detection data includes air quality data, water quality data, and noise data; Generate environmental monitoring results according to the urban environmental detection data; Formulate an environmental governance plan according to the environmental monitoring results; Generate pollution prevention and control measures according to the environmental governance plan, and send the pollution prevention and control measures to the visualization sub-module for visualization display.

[0016] In addition, to achieve the above object, the present invention further provides a digital twin smart city management device based on an AI model, where the digital twin smart city management device based on an AI model includes: a memory, a processor, and a digital twin smart city management program stored on the memory and executable on the processor. When the digital twin smart city management program is executed by the processor, it implements the steps of the digital twin smart city management method described in any one of the above.

[0017] In addition, to achieve the above object, the present invention further provides a computer-readable storage medium, on which a digital twin smart city management program based on an AI model is stored. When the digital twin smart city management program is executed by a processor, it implements the steps of the digital twin smart city management method described in any one of the above.

[0018] Compared with the prior art, the beneficial effects of the present invention are: Collect urban data of the target city, where the urban data includes real-time data collected by terminal devices installed in the city, static data and dynamic data obtained from business systems that have established a communication connection with the digital twin smart city management system, and citizen data collected from the network; construct a digital twin of the target city using the collected urban data and modeling techniques; deploy multiple AI models on the digital twin, where the multiple AI models are pre-trained using the urban data; the AI models include machine learning models and deep learning models, the machine learning models are used to process and analyze the obtained urban data to determine the real-time operation status and development needs of the city corresponding to the urban data; the deep learning models are used to process complex urban data and optimize urban management decisions; interact with the digital twin smart city management system through a visualization interface to obtain decision recommendation information analyzed by each of the multiple AI models. The embodiments of the present disclosure achieve digital modeling of the physical entities of the city, can monitor the operation status of the city in real time, predict the development trend of the city, optimize the resource allocation of the city, and improve the management efficiency and service level of the city. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of an embodiment of the digital twin smart city management system based on an AI model in an embodiment of the present invention; Figure 2 It is a schematic diagram of another embodiment of the digital twin smart city management system based on an AI model in an embodiment of the present invention; Figure 3Schematic diagram of an embodiment of the digital twin smart city management device based on the AI model in the embodiments of the present invention. Detailed implementation manners

[0020] The main object of the present invention is to provide a digital twin smart city management system based on the AI model; it realizes digital modeling of the physical entities of the city, can monitor the operation status of the city in real time, predict the development trend of the city, optimize the resource allocation of the city, and improve the management efficiency and service level of the city.

[0021] Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these process, method, product or device.

[0022] The digital twin smart city management system based on the AI model is an innovative system that uses artificial intelligence technology to comprehensively simulate and manage the smart city. It creates a digital twin of the city, presents elements such as urban infrastructure, resources, population, activities, etc. in the real world in a digital form, and uses the AI model for analysis, prediction and optimized management, so as to achieve the efficient, intelligent and sustainable development of urban operation.

[0023] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to Figure 1 , which is a schematic diagram of an embodiment of the digital twin smart city management system based on the AI model in the embodiments of the present invention. The digital twin smart city management system based on the AI model includes: Step 101, collect urban data of the target city, where the urban data includes real-time data collected by terminal devices installed in the city, static data and dynamic data obtained from business systems that establish communication connections with the digital twin smart city management system, and citizen data collected from the network. [[ID=ID=19]]

[0024] It can be understood that the execution subject of the present invention can be the digital twin smart city management system based on the AI model, or a terminal or a server. Specifically, it is not limited here. The embodiments of the present invention are described by taking the server as the execution subject as an example.

[0025] The terminal device may include environmental monitoring sensors and traffic flow sensors; the above-mentioned service system may include a government affairs office system and an urban planning database; among them, the data obtained in real time from the government affairs office system may be dynamic data; the data obtained from the urban planning database may be static data; the network may include social media, mobile applications, etc.; the above-mentioned citizen data may be data reflecting the lives and needs of citizens.

[0026] Optionally, the terminal devices installed in the city include traffic cameras, online monitoring devices, real-time monitoring systems, connecting to the urban Internet center, social media, connecting to public service facilities, and connecting to the urban Internet center; the urban data includes traffic data sets, infrastructure data sets, environmental data sets, security information sets, urban planning data sets, citizen interaction information sets, public service facility data sets, and market and social population data sets; the data acquisition module is used to collect urban data through the terminal devices installed in the city, including: the data acquisition module is used to collect traffic data sets through traffic cameras; collect infrastructure data sets through online monitoring devices; collect environmental data sets through online monitoring devices; collect security information sets through real-time monitoring systems; collect urban planning data sets through connecting to the urban Internet center; collect citizen interaction information sets through social media; collect public service facility data sets through connecting to public service facilities; collect market and social population data sets through connecting to the urban Internet center.

[0027] Step 102: Use the collected urban data and modeling techniques to construct a digital twin of the target city.

[0028] The above-mentioned modeling techniques may include 3D modeling, system dynamics modeling, etc. The digital twin is a highly accurate simulation of the real city, including not only the physical spatial layout of the city (such as buildings, roads, bridges, etc.), but also covering the operation logic of the city (such as traffic flow rules, energy consumption patterns, public service supply and demand relationships, etc.). It can be updated in real time to reflect the latest state of the city, enabling managers to conduct all-round observation and analysis of the city in a virtual environment.

[0029] Optionally, perform data cleaning on the urban data to obtain the cleaned data; classify the cleaned data according to the different types and uses of the urban data to obtain at least one data set; select corresponding data normalization methods for each data set according to the category of each data set in the at least one data set; perform data normalization processing on each data set respectively according to the corresponding data normalization methods to obtain at least one normalized data set; select corresponding modeling techniques for each normalized data set according to the category of each normalized data set; construct a digital twin of the target city according to each normalized data set and the corresponding modeling technique in the at least one normalized data set.

[0030] Perform data cleaning on the urban data to remove duplicate data, error data (such as outliers caused by sensor failures), and incomplete data records therein. For example, traffic flow sensors may generate some error data that significantly deviate from the normal range due to occasional communication failures. By setting reasonable thresholds and data verification rules, these abnormal data can be identified and eliminated. At the same time, for missing data, appropriate methods such as interpolation method and mean filling method can be used for supplementation to ensure the integrity of the data and provide an accurate data basis for subsequent modeling.

[0031] Classify the cleaned data according to the different types and uses of the urban data. As an example, it can be divided into infrastructure data (such as building information, road conditions, etc.), environmental data (such as air quality, water quality, etc.), population mobility data (such as people's travel trajectories, population density distributions, etc.), etc. Some key data can also be labeled to better identify and utilize these data during the modeling process. For example, label traffic congestion sections, marking information such as the time, location, and severity of congestion. These labeled information can be used as important references for training the model.

[0032] Data standardization methods can include Min-Max normalization (minimum-maximum normalization) method, Z-Score normalization (mean-standard deviation normalization) method, decimal scaling normalization method, logarithmic transformation normalization method, and Box-Cox transformation normalization method; as an example, in image processing, pixel values are usually between 0 and 255, but sometimes it is necessary to map them to the [0, 1] interval for the processing of certain algorithms, such as the input requirements of activation functions in neural networks. At this time, Min-Max normalization can be used to scale the original value of each pixel according to the minimum and maximum values of that pixel value in the image, so that it falls within the [0, 1] interval. When collecting data such as consumers' satisfaction scores and purchase frequencies for a certain product in market research, through Z-Score normalization, the data is converted into a form with a mean of 0 and a standard deviation of 1, which is convenient for subsequent statistical analysis and modeling operations. When evaluating the risk of a portfolio in the financial field, through Z-Score normalization, the return rate data of different assets can be unified to a standard scale, which is convenient for calculating risk indicators (such as Sharpe ratio, etc.) and comparing and evaluating the risks of different portfolios. In GIS, data such as terrain height and population density involved may also have large magnitude differences. For example, the terrain height may range from below sea level to several thousand meters of high mountains, and the population density may range from a few people per square kilometer to tens of thousands of people. Decimal scaling normalization can help process these data and make it more convenient for subsequent geographical analysis and modeling operations. When studying the income distribution of the social population, by taking the logarithm (such as natural logarithm or common logarithm) of the income data for normalization, the data distribution can be made closer to the normal distribution, which is convenient for statistical analysis, such as calculating statistical indicators such as the median and average of income, and for comparing the incomes between different regions and different groups.

[0033] Modeling techniques include regression analysis modeling techniques, decision tree modeling techniques, neural network modeling techniques, support vector machine (SVM) modeling techniques, clustering analysis modeling techniques, system dynamics modeling techniques, and Agent-Based Modeling techniques; regression analysis modeling techniques include linear regression and nonlinear regression; as an example, in the real estate market, linear regression can be used to analyze the relationship between housing prices (dependent variable) and independent variables such as housing area, house age, and surrounding supporting facilities, so as to predict the trend of housing prices; in the field of biomedicine, when studying the relationship between drug dosage (independent variable) and drug efficacy (dependent variable), a nonlinear relationship may be presented; in environmental science, analyze the relationship between pollutant concentration (dependent variable) and time (independent variable); decision tree modeling techniques are widely used in various classification scenarios, such as in bank credit risk assessment, according to the attribute characteristics of customers such as age, income, and credit record, judge whether the customer has credit risk (divided into two categories: with risk and without risk) through a decision tree model; in medical diagnosis, judge the types of diseases that patients may suffer from according to the symptoms, medical history, and examination results of patients. In neural network modeling techniques, CNN is widely used in the field of image recognition, such as face recognition systems, which can accurately identify the facial features of different people; in autonomous driving technology, identify objects such as road signs, vehicles, and pedestrians. RNN and its variants (such as long short-term memory network LSTM, gated recurrent unit GRU) perform well in speech recognition; SVM is mainly applied to classification scenarios, such as in spam filtering, according to the content characteristics of emails (such as keywords, sender addresses, etc.).

[0034] In marketing, consumers are divided into different groups through clustering analysis in order to formulate different marketing plans for different groups. In the field of image processing, clustering analysis can be used for image segmentation, dividing images according to similarities such as color and texture, so as to extract specific objects or regions in the images. System dynamics modeling techniques are used to analyze the operating mechanism and development trend of economic systems. By establishing a system dynamics model, predict the direction of economic development and provide a basis for formulating economic policies. In the study of ecological systems, by establishing a system dynamics model, study the dynamic change processes such as material cycling and energy flow in ecological systems. Agent-Based Modeling techniques are used in the field of social science to study social phenomena, such as the behavior patterns of crowds, the formation and development of social networks, etc. In economic systems, entities such as enterprises and consumers are regarded as individuals with autonomous behavior and decision-making abilities. By setting the behavior rules and interaction mechanisms of these entities, simulate the macro phenomena of the entire economic system.

[0035] Optionally, constructing the digital twin of the target city according to each of the at least one standardized data set and the corresponding modeling technique includes: constructing a digital twin of the physical space layout of the target city by using 3D modeling technology based on the standardized data set of infrastructure in the at least one standardized data set; constructing the operation logic of the target city by using system dynamics modeling technology and / or Agent-Based Modeling technology according to the at least one standardized data set; constructing a digital twin model of the target city based on the digital twin and the operation logic, and taking the digital twin model as the digital twin of the target city.

[0036] The standardized data set of infrastructure includes data of infrastructure such as buildings, roads, and bridges in the target city; the above-mentioned target city can be a city that needs to be intelligently managed; system dynamics modeling technology is mainly used to simulate the operation logic and dynamic process of the city. It describes the mutual relationship and dynamic change law among various elements in the city by establishing a series of differential equations or difference equations; Agent-Based Modeling technology simulates the behavior and phenomena of the city from the perspective of microscopic individuals. It regards each entity (such as residents, enterprises, vehicles, etc.) in the city as an individual with autonomous behavior and decision-making ability, and simulates the macroscopic phenomena of the entire city by setting the behavior rules and interaction mechanisms of these entities. By using 3D modeling technology, a physical space layout model of the city is constructed, including accurate three-dimensional models of infrastructure such as buildings, roads, and bridges. By using system dynamics modeling technology or Agent-Based Modeling technology, an operation logic model of the city is constructed according to the collected city data. This model should reflect the mutual relationship and dynamic change law among various elements in the city, such as traffic flow law, energy consumption pattern, public service supply-demand relationship, etc.

[0037] As an example, through 3D modeling software (such as Blender, 3ds Max, etc.), a three-dimensional virtual model of the city can be accurately created based on the data collected from infrastructure such as buildings, roads, and bridges; the external shape data of buildings obtained by lidar scanning can be combined with the internal structure information on architectural design drawings to construct a realistic building model in 3D modeling software. At the same time, a road network model of the city is constructed using satellite image data and road measurement data, and each building model is placed in the road network according to the actual location to form a complete physical space layout model of the city. When simulating the relationship between urban energy consumption and economic development, the mutual relationship equations between energy consumption and factors such as GDP growth, industrial structure adjustment, and population growth can be set. The parameters in these equations are determined by analyzing historical data, and then a model that can reflect the dynamic changes in urban energy consumption is constructed using system dynamics modeling technology, and the future trend of energy consumption can be predicted. When simulating urban traffic congestion, each vehicle can be regarded as an agent, and behavioral rules such as the driving speed of the vehicle and the rules for choosing the driving route (such as choosing the shortest path or the fastest path according to the traffic congestion situation) and the interaction mechanisms such as mutual avoidance and following between vehicles can be set. Through the simulation operation of a large number of agents, the formation, development, and alleviation processes of urban traffic congestion can be observed, providing a decision-making basis for traffic management.

[0038] Step 103, deploy multiple AI models on the digital twin, where the multiple AI models are pre-trained through the urban data; the AI models include machine learning models and deep learning models, and the machine learning models are used to process and analyze the obtained urban data to determine the real-time operation status and development needs of the city corresponding to the urban data; the deep learning models are used to process complex urban data and optimize urban management decisions.

[0039] Multiple AI models are used to deeply mine and analyze the data in the digital twin. Common AI models include machine learning models (such as decision trees, neural networks, etc.) for predictive analysis, deep learning models (such as convolutional neural networks for image recognition-related tasks, recurrent neural networks for processing sequential data, etc.) for processing complex urban data patterns, and reinforcement learning models for optimizing urban management decisions, etc. These AI models can realize various functions such as urban traffic flow prediction, environmental quality monitoring and prediction, energy consumption optimization, public safety event warning, etc., providing strong decision-making support for urban management.

[0040] Optionally, determine the correspondence between the categories of each standardized data set and each AI model among the multiple AI models; for different AI models deployed on the digital twin, determine the interface information for each AI model to receive data, where the interface information includes the format, type, and transmission frequency of the data transmitted by the corresponding interface; select a corresponding transmission protocol according to the data reception requirement information of each AI model, the corresponding interface information, and the deployment environment of the digital twin; build a data transmission channel from the data collection end to each AI model of the digital twin according to the selected transmission protocol; when the data arrives at the corresponding input interface of the digital twin through the transmission channel, parse the data according to the pre-determined interface information; integrate the parsed valid data into the corresponding AI model in the digital twin.

[0041] The categories of the data set can be divided into infrastructure data (such as building information, road conditions, etc.), environmental data (such as air quality, water quality, etc.), population flow data (such as people's travel trajectories, population density distribution, etc.), economic activity data (such as commercial sales, enterprise distribution, etc.), etc. Define the correspondence between each type of data and each functional module in the digital twin model. For example, infrastructure data mainly corresponds to the module in the digital twin model for presenting the urban physical space layout and infrastructure status; environmental data corresponds to the module for environmental monitoring and analysis; population flow data corresponds to the module for population distribution and flow simulation; economic activity data corresponds to the module related to economic development analysis.

[0042] The above requirement information can be the characteristics of the data, transmission requirements; as an example, for the traffic flow prediction module, its data input interface may require the transmission of traffic flow data, road condition data, etc. updated at regular intervals (such as every 5 minutes) in a specific format (such as JSON format); for the environmental monitoring module, its interface may require the transmission of real-time monitoring values of environmental data such as air quality and water quality, and the data format may be XML format. If the data volume is small and the real-time requirement is not particularly high, the HTTP protocol may be a suitable choice; if the data has real-time requirements and the data volume is relatively large, such as real-time traffic flow data, environmental monitoring data, etc., the MQTT or WebSocket protocol may be more suitable. When using the MQTT protocol to transmit real-time traffic flow data, an MQTT client needs to be installed at the data collection end (such as traffic flow sensors), and an MQTT broker server needs to be installed on the server where the digital twin model is located. By configuring the relevant parameters of the client and the broker server (such as topics, subscription relationships, etc.), a stable data transmission channel is established. If the data is transmitted in JSON format, the corresponding JSON parsing library is used to parse it into a format that can be processed by the model (for example, the json library in Python can parse a JSON string into data structures such as dictionaries or lists).

[0043] In addition, in order to ensure that the digital twin model can reflect the latest state of the city in real time, a real-time data update mechanism can be established. For example, for real-time traffic flow data, traffic flow sensors need to collect and send new data to the traffic flow prediction module of the digital twin model through the transmission channel at regular intervals (such as every 5 minutes). After receiving the new data, this module should immediately update the internal traffic flow status data for more accurate traffic flow prediction and analysis.

[0044] Optionally, in response to determining that the data format and characteristics provided by the digital twin match the AI model to be deployed, the urban data collected by the digital twin is further integrated to obtain the input data of the AI model; according to the application scenario information and expected function information of the digital twin, a target AI model is selected; a pre-evaluation is performed on the target AI model to determine the basic model information of the target AI model, where the basic model information includes accuracy, robustness, and computing resource requirements; according to the basic model information of each AI model in the multiple AI models, the overall requirement information of the computing resources for deploying the multiple AI models on the digital twin is determined; according to the overall requirement information, the device information of the target hardware device is determined; according to the pre-set model integration framework and the model parameters determined by each AI model in the multiple AI models during the model training process, the multiple AI models are deployed on the digital twin.

[0045] When it is determined that the data format and characteristics provided by the digital twin match the AI model to be deployed, the urban data collected by the digital twin is further integrated to obtain the input data of the AI model; different AI models may have different requirements for data. For example, some models may require a specific data normalization method, while some may have strict regulations on the dimensions, types, etc. of the data; urban data can be sensor data, business system data, etc.; for example, in the digital twin of a smart city, traffic flow sensor data, environmental monitoring data, urban infrastructure status data, etc. need to be reasonably integrated to meet the needs of different AI models such as traffic prediction and environmental analysis.

[0046] According to the application scenario and expected function of the digital twin, carefully select a suitable AI model; for example, for image recognition tasks in the digital twin (such as identifying specific objects or events in urban surveillance videos), a convolutional neural network (CNN) model may be selected; for the prediction of time series data (such as predicting the changing trend of urban energy consumption), models such as recurrent neural network (RNN) or its variants (such as long short-term memory network LSTM, gated recurrent unit GRU) may be considered. Conduct a pre-evaluation of the selected AI model, including aspects such as the accuracy, robustness, and computational resource requirements of the model.

[0047] Analyze the overall demand of multiple AI models to be deployed for computing resources (such as CPU, GPU, memory, etc.). Different AI models may consume different levels of hardware resources during the training and inference phases, especially deep learning models, which usually require a large amount of computing resources to ensure their efficient operation. Configure appropriate hardware devices according to the resource requirements. If it is deployed on a local server, appropriate CPU and GPU models and sufficient memory capacity can be selected according to the budget and performance requirements; if a cloud computing platform is used, the corresponding computing instance type (such as different types of EC2 instances on Amazon AWS) needs to be selected according to the model's requirements to ensure that there are sufficient hardware resources to support the simultaneous operation of multiple AI models.

[0048] The above model parameters include the weights, biases, learning rates, etc. of the model; common integration frameworks include TensorFlow Serving, ONNX Runtime, etc.; for example, TensorFlow Serving can efficiently deploy trained TensorFlow models (including multiple different types of models, such as classification models, prediction models, etc.); perform initialization operations on the models in the digital twin environment to make them in a runnable state. For example, when deploying a CNN model that has been trained on a large-scale dataset for image recognition tasks, it is necessary to load the trained weight file into memory and set the parameters of each layer according to the model's architecture so that it can receive the image data provided by the digital twin and perform recognition processing.

[0049] Step 104, interact with the digital twin smart city management system through the visualization interface to obtain the decision-making advice information analyzed by each of the multiple AI models.

[0050] City managers interact with the system through the visualization interface to obtain various results and suggestions analyzed by the AI models. They can make corresponding urban management decisions based on this information, such as adjusting the traffic signal timing, planning urban energy distribution, and allocating public service resources.

[0051] Optionally, the urban data includes traffic flow data, and the system further includes: real-time monitoring of the traffic flow data on urban roads; analyzing the causes and locations of traffic congestion according to the traffic flow data, and transmitting the traffic flow data to the visualization sub-module for visualization display; adjusting the working time of traffic lights according to the traffic flow data to optimize the traffic signal timing plan; controlling the working time of the traffic lights in the city according to the optimized traffic signal timing plan.

[0052] The system can real-time monitor the traffic flow on urban roads, analyze the causes and locations of traffic congestion. For example, using devices such as cameras and sensors to collect the vehicle driving data on the road, transmitting it to the digital twin model for real-time analysis and visualization display. According to the traffic flow data, intelligently adjust the time of traffic lights to optimize the traffic signal timing plan and improve the road traffic efficiency. For example, during the traffic peak period, automatically extend the green light time to reduce the vehicle waiting time; during the traffic low period, appropriately shorten the green light time to save energy.

[0053] The system can provide real-time traffic information and optimal travel route suggestions for travelers. For example, it can release road congestion conditions, traffic accident information, etc. to the public through mobile phone apps, electronic display screens, etc., guiding the public to choose the best travel route and avoid traffic congestion. It can also be combined with autonomous driving technology to achieve intelligent traffic guidance and coordinated operation of autonomous vehicles. For example, the digital twin system provides high-precision map information and real-time traffic conditions for autonomous vehicles, helping autonomous vehicles make more accurate decisions and improving driving safety and efficiency.

[0054] Optionally, the urban data includes urban environmental detection data, and the system further includes: monitoring the urban environment in real time, obtaining urban environmental detection data, and transmitting the urban environmental detection data to the visualization sub-module for visualization display, where the urban environmental detection data includes air quality data, water quality data, and noise data; generating an environmental monitoring result according to the urban environmental detection data; formulating an environmental governance plan according to the environmental monitoring result; generating pollution prevention and control measures according to the environmental governance plan, and sending the pollution prevention and control measures to the visualization sub-module for visualization display.

[0055] The system can monitor the urban environment in real time, including air quality, water quality, noise, etc. For example, environmental data is obtained through environmental monitoring sensors installed throughout the city and transmitted into the digital twin model for real-time analysis and visualization display. Analyze the environmental data to evaluate the changing trend and influencing factors of environmental quality. For example, by comparing and analyzing environmental data at different times and in different regions, identify the sources and transmission routes of environmental pollution, providing a scientific basis for environmental protection.

[0056] The system can intelligently formulate an environmental governance plan according to the environmental monitoring results. For example, when the air quality deteriorates, automatically initiate air pollution prevention and control measures, such as strengthening the supervision of industrial waste gas emissions and increasing the frequency of road sprinkling and dust suppression; when the water quality is polluted, promptly take water pollution control measures, such as strengthening sewage treatment and investigating pollution sources. Optimize and protect the urban ecosystem. For example, use digital twin technology to simulate the operation of the urban ecosystem, reasonably plan ecological spaces such as urban green spaces and wetlands, and improve the ecological environment quality of the city.

[0057] Continue to refer to Figure 2 , which is another schematic diagram of an embodiment of the digital twin smart city management system based on the AI model in the embodiment of the present invention. The digital twin smart city management system based on the AI model further includes: an urban planning module 201, an energy management module 202, and a public safety management module 203; The urban planning module includes a visualization sub-module and a construction management sub-module; The visualization sub-module is used to create a 3D model of the city in a virtual environment by using digital twin technology and the city data; the construction management sub-module is used to monitor the construction progress in real time, compare the differences between the planned progress and the actual progress, and adjust the construction arrangements.

[0058] By using digital twin technology, the visualization sub-module can create a 3D model of the city in a virtual environment and intuitively display the effects of different planning schemes. For example, it can simulate the impacts of the height and location of newly built buildings on the surrounding landscape and traffic, so as to select the optimal planning scheme. Through accurate modeling of the city's terrain, landforms, buildings, etc., it is possible to better analyze the land use situation and rationally layout the urban functional areas.

[0059] During the construction process, the construction management sub-module can monitor the construction progress in real time, compare the differences between the planned progress and the actual progress, and adjust the construction arrangements in a timely manner. For example, by installing sensors at the construction site, information such as the operating status of construction equipment and the work progress of workers can be obtained and fed back into the digital twin model to achieve refined management of the construction process. Monitor and give early warnings about the construction quality. For example, use sensors to detect the structural strength, deformation conditions, etc. of buildings. Once problems are found, notify the construction personnel in time for rectification to ensure the project quality.

[0060] The energy management module includes an energy detection and analysis sub-module and an energy optimization and scheduling sub-module; The energy detection and analysis sub-module is used to monitor the urban energy system in real time, transmit the detected energy data to the visualization sub-module for visual display, where the urban energy system includes a power system, a gas system, and a water supply system; analyze the detected energy data to determine the cause information of energy waste; and transmit the cause information to the energy optimization and scheduling sub-module; the energy optimization and scheduling sub-module is used to optimize the energy scheduling information according to the obtained energy demand information and energy supply situation.

[0061] The energy detection and analysis sub-module monitors the urban energy system in real time, including electricity, gas, water supply, etc. For example, by installing sensors on energy facilities, data on energy production, transmission, distribution, and consumption can be obtained and transmitted into the digital twin model for real-time analysis and visual display. Analyze the trends and patterns of energy consumption to identify the links and causes of energy waste. For example, by comparing and analyzing the energy consumption data in different regions and different time periods, areas or equipment with abnormal energy consumption can be found and repaired and optimized in a timely manner.

[0062] The energy optimization scheduling sub-module intelligently optimizes the scheduling and allocation of energy according to energy demand and supply conditions. For example, when the power supply is tight, it automatically adjusts the ratio of industrial electricity consumption and residential electricity consumption to give priority to ensuring residential electricity for daily use; when the gas supply is insufficient, it reasonably arranges the order of gas use to ensure the gas supply for important facilities and people's livelihood. Combined with renewable energy, it realizes the efficient utilization and sustainable development of energy. For example, it uses the digital twin system to predict the power generation of renewable energy such as solar energy and wind energy, and reasonably arranges the mixed use of renewable energy and traditional energy to improve energy utilization efficiency and reduce the impact on the environment.

[0063] The public safety management module includes a safety monitoring and early warning sub-module and an emergency rescue command sub-module; The safety monitoring and early warning sub-module is used to comprehensively monitor the urban public safety and obtain safety information in real time; based on the safety information, it generates event-related information and early warning information, and transmits the event-related information to the emergency rescue command sub-module, where the event-related information includes the location information of the event, the influence range information of the event, and the severity; the emergency rescue command sub-module is used to generate emergency plan information according to the event-related information; optimize the allocation and scheduling of rescue resources according to the emergency plan information.

[0064] The safety monitoring and early warning sub-module comprehensively monitors the urban public safety, including natural disasters such as fires, earthquakes, floods, and man-made disasters such as social security. For example, through devices such as sensors and cameras installed in every corner of the city, it obtains safety information in real time, transmits it to the digital twin model, and conducts real-time analysis and early warning. When a safety event occurs, the system can quickly locate the location of the event, analyze the influence range and severity of the event, automatically activate the emergency plan, and coordinate various departments for emergency disposal.

[0065] During the emergency rescue process, the above system can provide real-time on-site information and decision-making support for the commanders. For example, through virtual reality technology, the commanders can intuitively understand the situation at the accident scene in the digital twin model and formulate the optimal rescue plan; through the interconnection with rescue equipment, it realizes the real-time monitoring and command of the rescue process. Optimize the allocation and scheduling of rescue resources. For example, according to the type and severity of the accident, it automatically allocates rescue resources such as fire trucks, ambulances, and emergency rescue equipment to improve the rescue efficiency.

[0066] Next, the digital twin smart city management device based on the AI model in the embodiment of the present invention will be described in detail from the perspective of hardware processing.

[0067] Figure 3FIG. 0 is a schematic structural diagram of a digital twin smart city management device based on an AI model provided by an embodiment of the present invention. The digital twin smart city management device 300 based on the AI model may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) storing application programs 333 or data 332. Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of computer program operations in the digital twin smart city management device 300 based on the AI model. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of computer program operations in the storage media 330 on the digital twin smart city management device 300.

[0068] The digital twin smart city management device 300 based on the AI model may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 3 The shown structural diagram of the digital twin smart city management device based on the AI model does not limit the digital twin smart city management device based on the AI model, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0069] The present invention also provides a digital twin smart city management device based on an AI model, including: a memory and at least one processor, wherein a computer program is stored in the memory, and the memory and the at least one processor are interconnected by a line; the at least one processor invokes the computer program in the memory to cause the digital twin smart city management device based on the AI model to execute the steps in the above-mentioned digital twin smart city management method based on the AI model.

[0070] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. A computer program is stored in the computer-readable storage medium. When the computer program runs on a computer, the computer is caused to execute the steps of the digital twin smart city management method based on the AI model.

[0071] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0072] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several computer programs to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0073] The above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A digital twin smart city management system based on an AI model, characterized in that, Including: Collecting urban data of the target city, where the urban data includes real-time data collected by terminal devices installed in the city, static data, dynamic data obtained from business systems that have established communication connections with the digital twin smart city management system, and citizen data collected from the network; Using the collected urban data and modeling techniques to construct a digital twin of the target city; Deploying multiple AI models on the digital twin, where the multiple AI models are pre-trained through the urban data; the AI models include machine learning models and deep learning models. The machine learning models are used to process and analyze the obtained urban data to determine the real-time operating status and development needs of the city corresponding to the urban data; the deep learning models are used to process complex urban data and optimize urban management decisions; Interacting with the digital twin smart city management system through a visualization interface to obtain decision-making advice information analyzed by each AI model among the multiple AI models.

2. The digital twin smart city management system based on the AI model according to claim 1, characterized in that, The constructing the digital twin of the target city using the collected urban data and modeling techniques includes: Performing data cleaning processing on the urban data to obtain cleaned data; Classifying the cleaned data according to different types and uses of the urban data to obtain at least one data set; Selecting corresponding data standardization processing methods for each data set according to the category of each data set in the at least one data set; Performing data standardization processing on each data set respectively according to the corresponding data standardization processing methods to obtain at least one standardized data set; Selecting corresponding modeling techniques for each standardized data set according to the category of each standardized data set; Constructing the digital twin of the target city according to each standardized data set in the at least one standardized data set and the corresponding modeling techniques.

3. The digital twin smart city management system based on the AI model according to claim 2, wherein, The constructing the digital twin of the target city according to each standardized data set in the at least one standardized data set and the corresponding modeling techniques includes: Based on the standardized data set of the infrastructure in the at least one standardized data set, using 3D modeling techniques to construct a digital twin of the physical space layout of the target city; According to the at least one standardized data set, using system dynamics modeling techniques and / or Agent-Based Modeling techniques to construct the operating logic of the target city; Based on the digital twin and the operating logic, constructing a digital twin model of the target city and using the digital twin model as the digital twin of the target city.

4. The digital twin smart city management system based on the AI model according to claim 1, characterized in that, The deploying multiple AI models on the digital twin includes: Determining the correspondence between the category of each standardized data set and each AI model among the multiple AI models; For different AI models deployed on the digital twin, determining the interface information for each AI model to receive data, where the interface information includes the format, type, and transmission frequency of the data transmitted by the corresponding interface. Select a corresponding transmission protocol based on the requirement information for data received by each AI model, the corresponding interface information, and the deployment environment of the digital twin. Build a data transmission channel from the data acquisition end to each AI model of the digital twin according to the selected transmission protocol. When the data arrives at the corresponding input interface of the digital twin through the transmission channel, parse the data according to the pre-determined interface information. Integrate the parsed valid data into the corresponding AI model in the digital twin.

5. The digital twin smart city management system based on the AI model according to claim 1, wherein Deploying multiple AI models on the digital twin includes: In response to determining that the data format and features provided by the digital twin match the AI model to be deployed, further integrate the urban data collected by the digital twin to obtain the input data of the AI model. Select a target AI model according to the application scenario information and expected function information of the digital twin. Pre-evaluate the target AI model to determine the basic model information of the target AI model, where the basic model information includes accuracy, robustness, and computing resource requirements. Determine the overall requirement information of the computing resources for deploying multiple AI models on the digital twin according to the basic model information of each AI model in the multiple AI models. Determine the device information of the target hardware device according to the overall requirement information. Deploy the multiple AI models on the digital twin according to the pre-set model integration framework and the model parameters determined by each AI model in the multiple AI models during the model training process.

6. The digital twin smart city management system based on the AI model according to claim 1, characterized in that, The system further includes an urban planning module. The urban planning module includes a visualization sub-module and a construction management sub-module. The visualization sub-module is used to create a three-dimensional model of the city in a virtual environment by using digital twin technology and the urban data. The construction management sub-module is used to monitor the construction progress in real time, compare the difference between the planned progress and the actual progress, and adjust the construction arrangement.

7. An AI model-based digital twin smart city management system according to claim 6, characterized in that, The system further includes an energy management module. The energy management module includes an energy detection and analysis sub-module and an energy optimization scheduling sub-module. The energy detection and analysis sub-module is used to monitor the urban energy system in real time, transmit the detected energy data to the visualization sub-module for visual display, where the urban energy system includes a power system, a gas system, and a water supply system; analyze the detected energy data to determine the cause information of energy waste; and transmit the cause information to the energy optimization scheduling sub-module. The energy optimization scheduling sub-module is used to optimize the energy scheduling information according to the obtained energy demand information and energy supply situation.

8. The digital twin smart city management system based on the AI model according to claim 1, characterized in that, The system further includes a public safety management module. The public safety management module includes a safety monitoring and warning sub-module and an emergency rescue command sub-module. The safety monitoring and warning sub-module is used to monitor the urban public safety in all directions and obtain safety information in real time. Generate event-related information and early warning information based on the security information, and transmit the event-related information to the emergency rescue command sub-module, where the event-related information includes the location information of the event, the impact range information of the event, and the severity level. The emergency rescue command sub-module is used to generate emergency plan information according to the event-related information; optimize the allocation and scheduling of rescue resources according to the emergency plan information.

9. The digital twin smart city management system based on the AI model according to claim 6, wherein, The urban data includes traffic flow data, and the system further includes: Real-time monitor the traffic flow data of urban roads; Analyze the causes and locations of traffic congestion according to the traffic flow data, and transmit the traffic flow data to the visualization sub-module for visual display; Adjust the working time of traffic lights according to the traffic flow data, and optimize the traffic signal timing plan; Control the working time of traffic lights in the city according to the optimized traffic signal timing plan.

10. An AI model-based digital twin smart city management system according to claim 6, characterized in that, The urban data includes urban environmental detection data, and the system further includes: Monitor the urban environment in real time, obtain urban environmental detection data, and transmit the urban environmental detection data to the visualization sub-module for visual display, where the urban environmental detection data includes air quality data, water quality data, and noise data; Generate environmental monitoring results according to the urban environmental detection data; Formulate an environmental governance plan according to the environmental monitoring results; Generate pollution prevention and control measures according to the environmental governance plan, and send the pollution prevention and control measures to the visualization sub-module for visual display.

Images