A smart city global perception and collaborative scheduling system
The smart city full-domain perception and collaborative scheduling system, which integrates functional adaptation layers and AI intelligent agents, solves the problem of the separation between perception and scheduling, realizes real-time response and efficient scheduling, and improves the precision and security of urban management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHAANXI COVARIANCE INFORMATION TECHNOLOGY CO LTD
- Filing Date
- 2026-04-24
- Publication Date
- 2026-06-12
Abstract
Description
Technical Field
[0001] This invention relates to the field of smart city technology, and in particular to a smart city full-area perception and collaborative scheduling system. Background Technology
[0002] The construction of smart cities is currently moving into a deeper stage of "comprehensive perception, real-time response, and intelligent collaboration." However, existing technological solutions generally suffer from a disconnect between "perception" and "scheduling." On the one hand, sensing terminals deployed throughout the city, such as dedicated monitoring equipment integrating radar, cameras, and other sensors, can collect abundant urban operational data, but their function is limited to data collection and simple threshold alarms, unable to proactively trigger subsequent resource scheduling. On the other hand, the control terminals or systems responsible for resource scheduling rely on externally input data for their decisions, resulting in problems such as data delays, inconsistent formats, and information silos, causing scheduling instructions to lag behind real-time changes in urban events.
[0003] While some existing technologies attempt system integration, such as aggregating data from various systems through a data platform or interconnecting different devices through protocol converters, these solutions are mostly loosely coupled "data interfaces," failing to fundamentally resolve the functional and logical disconnect between dedicated sensing and scheduling devices. Sensing devices do not understand scheduling strategies, and scheduling devices lack awareness of the real-time status and subtle changes of sensing devices. This results in the entire urban management system lacking a complete closed-loop capability from "problem detection" to "problem resolution," hindering improvements in response efficiency and precision. Summary of the Invention
[0004] The present invention aims to provide a smart city full-area perception and collaborative scheduling system to solve the problems of fragmented perception and scheduling functions, poor coordination and delayed response in the existing technology.
[0005] To achieve the above objectives, the present invention provides the following technical solution:
[0006] A smart city full-area perception and collaborative scheduling system includes:
[0007] The functional fusion adaptation layer communicates with both the all-domain multi-dimensional perception terminal and the AI-enhanced intelligent resource collaborative scheduling terminal. The functional fusion adaptation layer includes:
[0008] The device protocol conversion module is used to convert the first communication protocol of the full-domain multi-dimensional perception terminal and the second communication protocol of the AI-enhanced intelligent resource collaborative scheduling terminal into a standard communication protocol to achieve interoperability at the protocol level.
[0009] The function mapping coupling module is used to build and store a linkage rule base of monitoring indicators and scheduling strategies, and to match the initial scheduling strategy from the linkage rule base based on the monitoring data uploaded by the full-domain multi-dimensional sensing terminal.
[0010] The all-domain perception and monitoring layer includes the all-domain multi-dimensional perception terminal, which is used to collect multi-dimensional raw monitoring data of urban operation;
[0011] The intelligent analysis and decision-making layer is equipped with a fusion AI analysis engine. The engine receives monitoring data processed by the functional fusion and adaptation layer, and generates risk levels and optimal scheduling schemes based on the built-in risk assessment model and scheduling optimization model.
[0012] The collaborative scheduling execution layer includes the AI-enhanced intelligent resource collaborative scheduling terminal. The terminal integrates an AI agent, which receives the optimal scheduling scheme, parses it into at least one executable instruction, sends it to the corresponding city execution device, and provides real-time feedback on the execution effect.
[0013] The function mapping coupling module also dynamically updates the linkage rule base based on the matching degree between the execution effect and the initial scheduling strategy.
[0014] Furthermore, the device protocol conversion module adopts a "protocol fingerprint recognition + lightweight adaptation algorithm" architecture to automatically identify and parse the LoRa or Wi-Fi protocol of the full-domain multi-dimensional sensing terminal, as well as the MQTT or HTTP protocol of the AI-enhanced intelligent resource collaborative scheduling terminal, and convert them into the GB / T39400-2020 standard protocol.
[0015] Furthermore, the functional mapping coupling module constructs the linkage rule base based on an improved BP neural network algorithm; the input layer of the improved BP neural network algorithm introduces urban operation status feature parameters, including holiday coefficients and peak period identifiers, with a weight adjustment step size of 0.01 to 0.05.
[0016] Furthermore, the full-domain multi-dimensional sensing terminal integrates millimeter-wave radar and high-definition camera, and adopts a spatiotemporal synchronization calibration algorithm, so that the time synchronization error between the millimeter-wave radar and the high-definition camera is ≤10ms and the spatial calibration accuracy is ≤0.5m.
[0017] Furthermore, the risk assessment model of the fusion AI analysis engine is a "random forest + LSTM" fusion model, in which random forest is used to process multi-dimensional static monitoring indicators, LSTM is used to capture the temporal change characteristics of indicators, and the output results of the two are weighted and fused to determine the risk level; the risk level includes: general risk, which is a single indicator exceeding the standard; high risk, which is two or more related indicators exceeding the standard; and major risk, which is a core indicator exceeding the standard and lasting for ≥5 minutes.
[0018] Furthermore, the AI agent includes:
[0019] The scene understanding unit is used to extract feature parameters from the monitoring data and match them with a preset scene feature library to identify the current urban operation scene;
[0020] The autonomous decision-making unit is used to generate a final scheduling scheme in a normal scenario based on the initial scheduling strategy matched by the functional mapping coupling module and real-time resource status data through a parameter optimization algorithm.
[0021] The reinforcement learning unit is used to collect the execution effect data of the scheduling scheme and iteratively optimize the decision model parameters of the autonomous decision-making unit based on the execution effect data.
[0022] Furthermore, it also includes an emergency response coordination layer, which has a built-in dynamic mapping algorithm between risk level and resource priority. When the risk level determined by the intelligent analysis and decision-making layer is a major risk, the emergency response coordination layer will raise the priority coefficient of emergency resources to above 0.9 and push information to the external emergency platform through an encrypted API interface.
[0023] Furthermore, it also includes a data security layer, which adopts a blockchain notarization mechanism to write the monitoring data collected by the multi-dimensional sensing terminal, the decision logs of the intelligent analysis and decision-making layer, the scheduling instructions issued by the collaborative scheduling and execution layer, and the received execution records as key data into the consortium blockchain in real time.
[0024] Furthermore, the data security layer optimizes the block generation logic, ensuring that the block generation time for critical data is ≤3s.
[0025] Furthermore, the functional fusion adaptation layer also includes an adaptive learning module, which continuously analyzes the matching degree between the monitoring data and the execution effect. When the matching degree is lower than a preset threshold, the functional mapping coupling module is triggered to optimize and adjust the mapping relationship in the linkage rule base.
[0026] Compared with the prior art, the present invention has the following beneficial effects:
[0027] 1. This invention, through a functional fusion adaptation layer, not only solves the problem of protocol heterogeneity between sensing terminals and scheduling terminals, but more importantly, through a functional mapping coupling module, establishes a dynamic correlation between monitoring indicators and scheduling strategies, enabling "sensing" to directly drive "scheduling" and forming a tight closed loop.
[0028] 2. The collaborative scheduling execution layer of this invention integrates an AI agent with scene understanding, autonomous decision-making and reinforcement learning capabilities, which can dynamically optimize the scheduling scheme based on real-time data and historical experience, making the scheduling instructions more accurate and efficient.
[0029] 3. The system of this invention forms a complete technology chain from data collection, protocol adaptation, intelligent analysis, autonomous decision-making to scheduling execution and effect feedback, which significantly shortens the response time from "event occurrence" to "completion of handling" and improves the level of precision in urban governance.
[0030] 4. This invention uses blockchain-based data security technology to ensure the immutability and traceability of critical data throughout its entire lifecycle, thus meeting the requirements of smart city management for data security and compliance.
[0031] 5. The reinforcement learning unit of the functional mapping coupling module and the AI agent of this invention enables the system to continuously optimize itself based on historical execution results, adapt to the dynamic changes in the city's operating status, and reduce manual intervention and maintenance costs. Detailed Implementation
[0032] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are merely illustrative of the present invention and are not intended to limit the scope of protection of the present invention.
[0033] Example
[0034] This embodiment provides a smart city full-domain perception and collaborative scheduling system, which is deployed in a typical urban environment.
[0035] System composition and connection relationships:
[0036] The system mainly consists of a functional integration and adaptation layer, a global perception and monitoring layer, an intelligent analysis and decision-making layer, a collaborative scheduling and execution layer, an emergency response and linkage layer, and a data security protection layer.
[0037] The city-wide perception and monitoring layer includes "city-wide multi-dimensional perception terminals" deployed throughout the city. These terminals integrate millimeter-wave radar and high-definition cameras, and through a built-in spatiotemporal synchronization calibration algorithm, ensure spatial and temporal consistency between the radar and camera when acquiring data on targets (such as vehicles and pedestrians), with errors controlled within 0.5m and 10ms respectively. It collects raw data in real time, including traffic flow, pedestrian density, and the status of municipal facilities.
[0038] Collaborative scheduling execution layer: This includes "AI-enhanced intelligent resource collaborative scheduling terminals" deployed in municipal, transportation, and security centers. These terminals communicate with various execution devices (such as traffic signals, patrol robots, and municipal maintenance work order systems).
[0039] Functional Integration and Adaptation Layer: Serving as a bridge connecting the two layers mentioned above, this layer is deployed within the city's data center. Its internal device protocol conversion module employs "protocol fingerprinting" technology to automatically identify the LoRa protocol from sensing terminals and the MQTT protocol from scheduling terminals. Through a "lightweight adaptation algorithm," it uniformly converts them to the GB / T39400-2020 standard protocol within 50ms, achieving seamless data integration. Its internal functional mapping and coupling module incorporates a "monitoring data - risk level - scheduling scheme" linkage rule base trained using an improved BP neural network. For example, when receiving monitoring data indicating "traffic flow on a certain road segment > 100 vehicles / minute," this module will initially match an initial strategy of "activating the traffic diversion plan around this road segment."
[0040] System workflow:
[0041] 1. Data Acquisition and Protocol Adaptation: The multi-dimensional sensing terminal collects raw data indicating that "the population density in business district A exceeds 3 people / ㎡" and transmits it via the LoRa protocol. The device protocol conversion module of the function fusion and adaptation layer receives this data, completes the protocol conversion, and simultaneously sends the standard format data to the function mapping coupling module and the intelligent analysis and decision-making layer.
[0042] 2. Initial Matching and In-Depth Analysis: Based on the indicator of "excessive personnel density," the functional mapping coupling module matches the initial strategy of "initiating personnel evacuation." Simultaneously, the fusion AI analysis engine of the intelligent analysis and decision-making layer is activated. Its "Random Forest" sub-model processes static indicators such as current personnel density, area type (business district), and time (Friday evening rush hour), while the "LSTM" sub-model analyzes the growth trend of personnel density over the past 15 minutes. Both analyses combined determine this as "high-risk."
[0043] 3. Generation and optimization of scheduling scheme: The intelligent analysis decision-making layer's scheduling optimization model, combined with real-time resource load coefficients (such as 3 patrol robots available and 2 auxiliary police officers on duty in the vicinity), generates an optimal scheduling scheme through a multi-objective optimization algorithm, which includes "scheduling 2 patrol robots to the entrance of business district A for voice guidance and notifying surrounding bus stations to add backup bus services".
[0044] 4. Autonomous Decision-Making and Command Execution: The AI agent in the collaborative scheduling execution layer receives the plan. Its scenario understanding unit identifies the current scenario as "peak-hour traffic management in a commercial district." Based on the robots' real-time locations, the autonomous decision-making unit refines the command "dispatch two patrol robots" into specific instructions such as "Robot A moves from its current location along route X to entrance 1, and Robot B moves from its current location along route Y to entrance 2," and issues these instructions through the scheduling terminal. The robots begin to move and provide real-time feedback on their location and status.
[0045] 5. Effect Feedback and Reinforcement Learning: During execution, the sensing terminal continuously monitors changes in pedestrian density in the business district. When it is found that the crowd control effect is not as expected, the reinforcement learning unit will record this situation and use "current solution - actual effect" as a set of data to optimize the decision-making logic of the autonomous decision-making unit in the future. For example, in the future, it may prioritize increasing the number of manual crowd control personnel rather than relying solely on robots.
[0046] 6. Emergency Response and Safety Assurance: If a sudden drop in pipeline pressure occurs in a certain area and lasts for more than 5 minutes, the intelligent analysis and decision-making layer determines it as a "major risk." The emergency response linkage layer immediately triggers an emergency response, setting the priority coefficient of emergency resources such as valve closure and repair to 0.95, and pushing early warning information containing the risk location, situation, and suggested handling plan to the Emergency Management Bureau through an encrypted API interface. Throughout the entire process, from the raw data of the sensing terminals and the judgment logs of the analysis and decision-making layer to every instruction issued by the dispatch terminal, everything is written to the consortium blockchain in real time by the data security assurance layer. The block generation time is less than 3 seconds, ensuring the traceability and immutability of the entire event handling process.
[0047] Implementation effect verification
[0048] After pilot application in a part of a medium-sized city, compared with the traditional standalone system, the average response time for municipal facility failures was reduced from 60 minutes to 25 minutes, and the linkage response time for major emergency events was reduced from 40 minutes to 16 minutes. The system demonstrated extremely high collaborative efficiency and intelligence.
[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A smart city full-area perception and collaborative scheduling system, characterized in that, include: The functional fusion adaptation layer communicates with both the all-domain multi-dimensional perception terminal and the AI-enhanced intelligent resource collaborative scheduling terminal. The functional fusion adaptation layer includes: The device protocol conversion module is used to convert the first communication protocol of the full-domain multi-dimensional perception terminal and the second communication protocol of the AI-enhanced intelligent resource collaborative scheduling terminal into a standard communication protocol to achieve interoperability at the protocol level. The function mapping coupling module is used to build and store a linkage rule base of monitoring indicators and scheduling strategies, and to match the initial scheduling strategy from the linkage rule base based on the monitoring data uploaded by the full-domain multi-dimensional sensing terminal. The full-domain perception and monitoring layer includes the full-domain multi-dimensional perception terminal, which is used to collect multi-dimensional raw monitoring data of urban operation; The intelligent analysis and decision-making layer is equipped with a fusion AI analysis engine. The engine receives monitoring data processed by the functional fusion and adaptation layer, and generates risk levels and optimal scheduling schemes based on the built-in risk assessment model and scheduling optimization model. The collaborative scheduling execution layer includes the AI-enhanced intelligent resource collaborative scheduling terminal. The terminal integrates an AI agent, which receives the optimal scheduling scheme, parses it into at least one executable instruction, sends it to the corresponding city execution device, and provides real-time feedback on the execution effect. The function mapping coupling module also dynamically updates the linkage rule base based on the matching degree between the execution effect and the initial scheduling strategy.
2. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The device protocol conversion module adopts a "protocol fingerprint recognition + lightweight adaptation algorithm" architecture to automatically identify and parse the LoRa or Wi-Fi protocol of the full-domain multi-dimensional sensing terminal, as well as the MQTT or HTTP protocol of the AI-enhanced intelligent resource collaborative scheduling terminal, and convert them into the GB / T39400-2020 standard protocol.
3. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The functional mapping coupling module constructs the linkage rule base based on an improved BP neural network algorithm; the input layer of the improved BP neural network algorithm introduces urban operation status feature parameters, including holiday coefficients and peak period identifiers, with a weight adjustment step size of 0.01 to 0.
05.
4. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The full-domain multi-dimensional sensing terminal integrates millimeter-wave radar and high-definition camera, and adopts a spatiotemporal synchronization calibration algorithm to ensure that the time synchronization error between the millimeter-wave radar and the high-definition camera is ≤10ms and the spatial calibration accuracy is ≤0.5m.
5. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The risk assessment model of the fusion AI analysis engine is a "random forest + LSTM" fusion model. Random forest is used to process multi-dimensional static monitoring indicators, and LSTM is used to capture the temporal change characteristics of the indicators. The output results of the two are weighted and fused to determine the risk level. The risk level includes: general risk, which is a single indicator exceeding the standard; high risk, which is two or more related indicators exceeding the standard; and major risk, which is a core indicator exceeding the standard and lasting for ≥5 minutes.
6. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The AI agent includes: The scene understanding unit is used to extract feature parameters from the monitoring data and match them with a preset scene feature library to identify the current urban operation scene; The autonomous decision-making unit is used to generate a final scheduling scheme based on the initial scheduling strategy matched by the functional mapping coupling module and real-time resource status data in a normal scenario, through a parameter optimization algorithm. The reinforcement learning unit is used to collect the execution effect data of the scheduling scheme and iteratively optimize the decision model parameters of the autonomous decision-making unit based on the execution effect data.
7. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, It also includes an emergency response coordination layer, which has a built-in dynamic mapping algorithm between risk level and resource priority. When the risk level determined by the intelligent analysis and decision-making layer is a major risk, the emergency response coordination layer will raise the priority coefficient of emergency resources to above 0.9 and push information to the external emergency platform through an encrypted API interface.
8. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, It also includes a data security layer, which uses a blockchain notarization mechanism to write the monitoring data collected by the multi-dimensional sensing terminal, the decision logs of the intelligent analysis and decision-making layer, the scheduling instructions issued by the collaborative scheduling and execution layer, and the received execution records into the consortium blockchain in real time as key data.
9. The smart city full-domain perception and collaborative scheduling system according to claim 8, characterized in that, The data security layer optimizes the block generation logic, ensuring that the block generation time for critical data is ≤3s.
10. The smart city full-domain perception and collaborative scheduling system according to claim 1, characterized in that, The functional fusion adaptation layer also includes an adaptive learning module, which continuously analyzes the matching degree between the monitoring data and the execution effect. When the matching degree is lower than a preset threshold, the functional mapping coupling module is triggered to optimize and adjust the mapping relationship in the linkage rule base.