Rail transit intelligent control system and method based on Internet of Things

Abstract

The invention relates to the field of rail transit control systems, and discloses a rail transit intelligent control system and method based on the Internet of Things, and the system comprises a sensing layer which is integrated with a visual sensor, a thermal imaging sensor, a sound sensor, a pressure sensor, a temperature and humidity sensor, a CO2 sensor and other sensors, the sensing layer collects various data of the rail transit system in real time, and the sensing layer performs joint optimization processing on sensor data by adopting deep learning and reinforcement learning algorithms through an adaptive data fusion method so as to realize high-precision real-time monitoring, data quality screening and anomaly detection; and the network layer comprises a wireless communication module and adopts an intelligent routing mechanism based on real-time network load prediction. A self-adaptive data fusion method is adopted by the sensing layer, and the sensing capability of the rail transit system is greatly improved through intelligent data processing, so that the system can realize real-time and accurate response to a complex traffic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of rail transit control systems, and in particular to an intelligent rail transit control system and method based on the Internet of Things. Background Art

[0002] Rail transit system is an important part of modern urban transportation network. With the continuous advancement of urbanization, the scale and complexity of rail transit operation are increasing. Traditional rail transit system relies on manual dispatching, manual fault detection and fixed operation process, which is difficult to cope with the growing passenger flow and changing operation environment. In order to improve the efficiency, reliability and safety of rail transit, modern rail transit system is gradually developing in the direction of intelligence and automation, especially the application of intelligent control system based on Internet of Things (IoT) technology, which has become the key to improving the comprehensive efficiency of rail transit system.

[0003] With the rapid development of technologies such as the Internet of Things, artificial intelligence, and edge computing, the intelligence level of rail transit systems has also been greatly improved. Existing rail transit intelligent control systems usually rely on the support of a large number of sensors and real-time data streams to achieve comprehensive monitoring of trains, tracks, environments, and passenger behaviors. At the same time, scheduling algorithms and fault prediction systems based on artificial intelligence can optimize scheduling plans, predict equipment failures, and improve system safety in real time through data analysis. However, although the current system has made some progress in some areas, there are still some technical bottlenecks that affect the real-time, accuracy, and adaptability of the overall system.

[0004] At present, most of the existing rail transit intelligent control systems focus on a single data processing and decision-making optimization module, lacking a fully integrated system architecture. Although the existing system can monitor the status of trains, tracks and the environment through sensors and perform certain scheduling optimization, the data transmission relies on the central processing system and has a large data transmission delay, which makes it impossible to achieve immediate response to tasks such as fault prediction, scheduling optimization and safety warning. Especially in a complex operating environment, the system's adaptability to dynamic changes and real-time response capabilities still have a lot of room for improvement. In addition, the accuracy and timeliness of the fault prediction and safety warning systems also need to be further improved to reduce the probability of equipment failures and safety accidents, and ensure the safe and efficient operation of the rail transit system. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an Internet of Things-based rail transit intelligent control system and method thereof, which solves the problem that the prior art rail transit intelligent control systems are mostly concentrated on a single data processing and decision optimization module and lack a fully integrated system architecture.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent rail transit control system based on the Internet of Things, the system comprising:

[0007] The perception layer collects various data of the rail transit system in real time by integrating multiple sensors such as visual sensors, thermal imaging sensors, sound sensors, pressure sensors, temperature and humidity sensors, and CO2 sensors. The perception layer uses an adaptive data fusion method and adopts deep learning and reinforcement learning algorithms to jointly optimize and process sensor data to achieve high-precision real-time monitoring, data quality screening, and anomaly detection;

[0008] The network layer, including the wireless communication module, adopts an intelligent routing selection mechanism based on real-time network load prediction, dynamically selecting the data transmission path according to the real-time network status and load to ensure the stability and low latency of data transmission;

[0009] The application layer includes an intelligent dispatching module, a fault prediction module and a safety warning module. The intelligent dispatching module dynamically adjusts the train dispatching plan according to real-time data by adopting a multi-objective optimization algorithm. The fault prediction module jointly analyzes historical data and real-time data through a deep neural network and a self-supervised learning algorithm to predict equipment failures and generate warning signals. The safety warning module determines potential safety hazards based on multiple sensor data and triggers automatic response measures.

[0010] Edge computing nodes are used for real-time computing and analysis of local data, generating scheduling instructions, fault warnings or safety warnings, and reducing data transmission delays and optimizing local decision responses through edge intelligent decision-making algorithms;

[0011] The management platform is used to centrally display all sensor data, scheduling plans, fault status, and early warning information, and provide a manual intervention control interface to allow managers to make necessary adjustments to the system. The platform further uses an adaptive optimization algorithm to adjust the overall performance of the system.

[0012] Preferably, the sensors of the perception layer adopt heterogeneous multimodal sensor data fusion technology to realize the joint processing of multiple sensor data, and adopt an adaptive weighted fusion algorithm so that the system can automatically adjust the sensor weights under different environmental conditions and optimize data accuracy.

[0013] Preferably, the network layer predicts network traffic and bandwidth requirements in real time through an artificial intelligence-based network load prediction model, and adopts an intelligent routing dynamic optimization algorithm to reduce network congestion and delay, thereby ensuring high-reliability data transmission.

[0014] Preferably, the intelligent scheduling module adopts a multi-objective optimization algorithm based on a combination of genetic algorithm and particle swarm optimization algorithm to perform scheduling calculations based on real-time train position, speed, track condition, traffic flow and system load data, generate a train operation plan, and dynamically optimize the train's operation path, speed, stop time, etc., to further improve the response speed and accuracy of train scheduling, and reduce delays, energy consumption and congestion.

[0015] Preferably, the fault prediction module adopts a machine learning method based on a combination of deep neural networks and time series models to accurately predict equipment failures through a comprehensive analysis of historical data, real-time data, and equipment operating parameters. The algorithm is continuously optimized through reinforcement learning to improve prediction accuracy and reduce the occurrence of failures.

[0016] Preferably, the safety warning module fuses data from infrared thermal imaging sensors, smoke sensors, vibration sensors, acceleration sensors and environmental gas sensors, and adopts abnormal pattern recognition technology based on deep learning. When fire, derailment, abnormal vibration or other safety incidents are detected, an alarm is immediately issued, and multi-level safety response measures are triggered through an automated control system, including automatic parking, evacuation instructions and warning broadcasts.

[0017] Preferably, the edge computing node performs real-time calculations on a large amount of sensor data through local data processing and aggregation technology, and makes scheduling optimization, fault detection or safety alarm decisions based on the calculation results. The decisions are executed locally by the edge computing node to reduce data transmission delays.

[0018] Preferably, the system further includes a management platform, which is used to centrally display all sensor data, scheduling plans, fault status and safety warning information, and provide a manual intervention control interface to allow managers to make necessary adjustments to the system.

[0019] The rail transit intelligent control method based on the Internet of Things includes the following steps:

[0020] Step 1: Use multiple heterogeneous sensors to collect various data of the rail transit system in real time, including train status, track environment, car environment and passenger behavior data, and use adaptive data fusion algorithm for data preprocessing;

[0021] Step 2: The collected data is transmitted to the edge computing node for preliminary processing through a wireless communication protocol. The data preprocessing includes denoising, standardization, anomaly detection, etc., and the transmission path is optimized through an intelligent routing selection algorithm based on deep learning;

[0022] Step 3: Perform fusion analysis on the processed data at the application layer, predict equipment failures based on historical data and real-time data, generate scheduling optimization plans, and issue safety warnings. The failure prediction is performed by combining deep learning with time series analysis algorithms.

[0023] Step 4: Generate the train's dispatch path, speed and stop time based on the data analysis results, and dynamically optimize the dispatch plan through the intelligent dispatch module and send it to the train for execution;

[0024] Step 5: When a potential fault or safety hazard is detected, emergency response measures are immediately triggered, and decisions are locally executed through edge computing nodes to adjust train operations or stop operations and issue safety alerts.

[0025] Preferably, the method adopts a scheduling optimization algorithm based on a combination of a deep neural network and a particle swarm optimization algorithm. The method also adopts an equipment fault prediction method that combines a convolutional neural network and a support vector machine. The method also uses a self-supervised learning algorithm to perform real-time data analysis and timely adjusts the scheduling, control and alarm strategies according to the prediction results. The algorithm is continuously optimized through reinforcement learning to improve the system prediction accuracy and response speed.

[0026] The present invention provides an intelligent rail transit control system and method based on the Internet of Things.

[0027] Beneficial effects:

[0028] 1. The present invention integrates multiple sensors such as visual sensors, thermal imaging sensors, sound sensors, pressure sensors, temperature and humidity sensors, CO2 sensors, etc. through the perception layer, which can comprehensively monitor the train, track, carriage and passenger behavior data. The perception layer adopts an adaptive data fusion method, combined with deep learning and reinforcement learning algorithms to optimize the processing of sensor data, ensuring high-precision real-time monitoring, data quality screening and anomaly detection. This intelligent data processing greatly improves the perception ability of the rail transit system, enabling the system to achieve real-time and accurate response to complex traffic environments.

[0029] 2. The intelligent dispatching module of the present invention dynamically adjusts the train dispatching plan through a multi-objective optimization algorithm and realizes intelligent dispatching of train operation according to real-time data. The system can automatically calculate and optimize the running path, speed, stop time, etc. of the train, reduce delays, energy consumption, and alleviate traffic congestion, thereby improving the overall operation efficiency of the system.

[0030] 3. The present invention uses an integrated safety warning module, so the system can quickly detect potential safety hazards such as fire, derailment, abnormal vibration, etc., and automatically sound an alarm when a safety incident occurs, triggering multi-level safety responses such as automatic parking, evacuation instructions, and warning broadcasts. This enables the rail transit system to quickly take emergency measures when sudden safety incidents occur, ensuring the safety of passengers and equipment and improving the safety of the system.

[0031] 4. The local data processing capability of the edge computing node of the present invention significantly reduces data transmission delay. The edge computing node can not only process a large amount of sensor data in real time, but also quickly make scheduling optimization, fault detection and safety alarm decisions based on the processing results. This local execution method greatly improves the speed and real-time response of the system, enabling the rail transit system to respond to complex dynamic environments and emergencies in a timely manner.

[0032] 5. The fault prediction module of the present invention combines deep neural networks, time series analysis and reinforcement learning algorithms to comprehensively analyze the equipment operation status, sensor data and historical fault data, accurately predict the potential failure of the equipment and issue early warning signals in a timely manner. Through the reinforcement learning optimization algorithm, the system can continuously improve the accuracy of prediction, reduce the occurrence rate of failures, reduce the risk of equipment downtime, and improve the reliability and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flow chart of the method of the present invention;

[0034] Figure 2 It is the main framework diagram of the present invention;

[0035] Figure 3 It is a schematic diagram of the network layer process of the present invention;

[0036] Figure 4 It is a schematic diagram of the edge computing node-to-point process of the present invention;

[0037] Figure 5 It is a schematic diagram of the application layer process of the present invention;

[0038] Figure 6 It is a schematic diagram of the management platform flow of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Embodiment 1:

[0041] Please see attached Figure 2 -Attached Figure 6 The embodiment of the present invention provides a rail transit intelligent control system based on the Internet of Things, the system comprising:

[0042] The perception layer collects various data of the rail transit system in real time by integrating multiple sensors such as visual sensors, thermal imaging sensors, sound sensors, pressure sensors, temperature and humidity sensors, and CO2 sensors. The perception layer uses adaptive data fusion methods and uses deep learning and reinforcement learning algorithms to jointly optimize and process sensor data to achieve high-precision real-time monitoring, data quality screening, and anomaly detection.

[0043] The network layer, including the wireless communication module, adopts an intelligent routing selection mechanism based on real-time network load prediction, dynamically selecting the data transmission path according to the real-time network status and load to ensure the stability and low latency of data transmission;

[0044] The application layer includes the intelligent dispatching module, the fault prediction module and the safety warning module. The intelligent dispatching module dynamically adjusts the train dispatching plan according to real-time data by adopting a multi-objective optimization algorithm. The fault prediction module jointly analyzes historical data and real-time data through deep neural networks and self-supervised learning algorithms to predict equipment failures and generate warning signals. The safety warning module determines potential safety hazards based on multiple sensor data and triggers automatic response measures.

[0045] Edge computing nodes are used for real-time computing and analysis of local data, generating scheduling instructions, fault warnings or safety warnings, and reducing data transmission delays and optimizing local decision responses through edge intelligent decision-making algorithms;

[0046] The management platform is used to centrally display all sensor data, scheduling plans, fault status, and early warning information, and provides a manual intervention control interface, allowing managers to make necessary adjustments to the system. The platform further uses an adaptive optimization algorithm to adjust the overall performance of the system.

[0047] In one embodiment, the perception layer: data collection and fusion

[0048] The perception layer is the data collection layer of the rail transit intelligent control system, responsible for collecting various types of data in the rail transit system in real time. This layer uses multiple heterogeneous sensors, sensor nodes and data acquisition devices to achieve comprehensive monitoring of train status, track environment, car environment, passenger behavior, etc.

[0049] Combine environmental sensor data with big data analysis for energy-saving control. Based on the environmental monitoring data of the station hall and platform, the system analyzes temperature, humidity and passenger flow data to intelligently adjust the use status of the station's air conditioning, heating and lighting equipment to achieve energy saving and consumption reduction. If the temperature is suitable, the air conditioning will automatically reduce the power; if the passenger flow is small, the station lighting will automatically dim.

[0050] A big data prediction model based on ambient temperature and passenger flow is introduced, combined with data from temperature and humidity sensors, and the equipment status is intelligently adjusted through algorithms to reduce unnecessary energy consumption.

[0051] Sensor type and arrangement:

[0052] Visual sensor: used to monitor the external environment of the rail transit system, detect the running status of the train, obstacles around the station and the condition of the road, and assist the dispatching system in making safety judgments.

[0053] Thermal imaging sensors: used to detect temperature changes in train equipment, tracks and station facilities, and monitor whether the equipment is overheating or has potential fire hazards.

[0054] Sound sensor: detects noise changes when the train is running and helps diagnose possible mechanical failures, such as axle abnormalities.

[0055] Pressure sensor: installed on the track, it detects the pressure of the train on the track in real time, helping to analyze the train load, track pressure and whether there is any imbalance or abnormality.

[0056] Temperature and humidity sensors: Installed in the carriages and track areas, they monitor environmental conditions in real time to ensure that the train and station environments meet comfort requirements.

[0057] CO2 sensor: used to monitor the air quality in the car and detect whether the oxygen concentration is normal in order to adjust the ventilation and air-conditioning system in the car.

[0058] Other sensors: such as vibration sensors, acceleration sensors, smoke sensors, GPS positioning sensors, etc., are used to monitor the running status of trains and the safety of tracks in real time.

[0059] Data collection and preprocessing:

[0060] Data collection is carried out synchronously through various sensors and sensor nodes to collect real-time status data of trains, tracks, carriages, passengers, etc. The data is transmitted to the network layer through wireless communication modules (such as Zigbee, NB-I oT).

[0061] Data preprocessing includes steps such as denoising, standardization, and anomaly detection. The adaptive data fusion algorithm optimizes the sensor data fusion process through deep learning and reinforcement learning, and dynamically adjusts the weight and processing method of sensor data to cope with data changes in different environments.

[0062] Adaptive weighted fusion algorithm: adjusts the weight of each sensor data in real time according to the reliability of the sensor and environmental factors. For example, in bad weather or complex environment, the accuracy of some sensors may decrease. The system will automatically adjust the weight of the fusion algorithm to ensure the accuracy of data processing.

[0063] Data quality screening and anomaly detection

[0064] Based on deep learning and reinforcement learning algorithms, the system can filter out high-quality data from sensor data and perform real-time anomaly detection. By training models and optimizing algorithms, the system can detect abnormal conditions in tracks, trains, and carriages in real time, such as equipment failures and environmental anomalies.

[0065] Network layer: data transmission and intelligent routing

[0066] The main task of the network layer is to ensure that the data collected by the perception layer can be stably and low-latency transmitted to the edge computing node for further processing. This layer uses wireless communication technology and intelligent routing optimization algorithms to ensure that the system's data transmission is not affected by network fluctuations.

[0067] Data transmission:

[0068] Adopt 4G / 5G communication protocol to transmit large-scale data, ensure real-time data transmission between trains and stations, and adapt to high-frequency and low-latency network requirements.

[0069] Communication technologies such as Wi-Fi and Zigbee are used for low-power data transmission inside stations and carriages, ensuring that the system can operate efficiently while avoiding waste of resources.

[0070] LoRaWAN communication technology is used in remote areas or devices that cannot access traditional networks to ensure full coverage of the system.

[0071] Intelligent routing selection mechanism:

[0072] The network layer uses an AI-based network load prediction model to predict network traffic and bandwidth requirements in real time. Based on factors such as network status, routing load, and bandwidth requirements, the system can dynamically select the best data transmission path.

[0073] Intelligent routing optimization algorithm: Combines real-time network traffic, bandwidth requirements and delay prediction, and uses deep learning to optimize routing selection strategies to reduce network congestion and ensure efficient and stable data transmission.

[0074] Reliability of data transmission:

[0075] The network layer uses prediction algorithms to achieve balanced distribution of data traffic, reduce data congestion, and ensure high system reliability and low-latency data transmission.

[0076] Application layer: intelligent scheduling, fault prediction and safety warning

[0077] The application layer is the core of the rail transit intelligent control system, responsible for real-time data analysis, intelligent scheduling, fault prediction and safety warning.

[0078] Through the interface between the signal control system and the station integrated monitoring system, the system can monitor the arrival status of trains in real time and automatically adjust the brightness of station lights according to the arrival and departure of trains. For example, when a train arrives, the station platform lights are automatically brightened; after the train leaves, the lights are automatically dimmed, thus achieving energy saving.

[0079] Specific operation steps: In the application layer or management platform, the train and station status are identified through intelligent algorithms, and data interaction and dispatch control are carried out in the cloud system to ensure that the lighting status can be dynamically adjusted according to demand to avoid energy waste.

[0080] Intelligent scheduling module:

[0081] A multi-objective optimization algorithm based on a combination of genetic algorithm and particle swarm optimization algorithm is used to dynamically calculate the train scheduling plan based on real-time data (train position, speed, track condition, traffic flow, etc.).

[0082] By analyzing the train operation status and track conditions in real time, the train operation path, speed, stop time, etc. are optimized to ensure the accuracy and response speed of train scheduling and reduce delays and energy consumption.

[0083] In the intelligent dispatching module, the real-time passenger flow forecast and historical data analysis are combined to adjust the departure interval and vehicle dispatching plan according to the passenger flow changes. In periods with low passenger flow, the departure frequency and running speed of trains are reduced to reduce energy consumption; while in periods with high passenger flow, the system automatically optimizes the train dispatching plan to improve transportation efficiency.

[0084] Algorithm support: Through multi-objective optimization algorithms and deep learning models, the intelligent dispatching system can dynamically adjust the train departure intervals, stop sites and dispatching priorities to ensure maximum energy utilization while meeting passenger flow demand.

[0085] Fault prediction module:

[0086] A machine learning method based on a combination of deep neural networks and time series analysis is used to analyze equipment operating status, sensor data, and historical fault data to accurately predict equipment failures and generate early warning signals.

[0087] Reinforcement learning continuously optimizes the fault prediction model, enhances the system's ability to foresee potential faults, reduces the occurrence of faults, and improves system reliability.

[0088] Security warning module:

[0089] Integrate multiple sensor data (such as infrared thermal imaging, smoke, vibration, acceleration sensors, etc.), and use deep learning-based abnormal pattern recognition technology to detect safety hazards such as fire, derailment, abnormal vibration, etc. in real time.

[0090] Once a safety hazard is detected, the system immediately sounds an alarm and triggers safety response measures such as automatic parking, evacuation instructions and warning broadcasts to ensure the safety of passengers.

[0091] Edge computing nodes: local data processing and optimization

[0092] Edge computing nodes are used for real-time calculation and analysis of local data, reducing data transmission delays and making optimization decisions.

[0093] Data processing and calculation:

[0094] Edge computing nodes perform real-time calculations and processing on large amounts of data collected by sensors, such as scheduling instruction generation, fault detection, data aggregation, etc., to reduce pressure on network bandwidth.

[0095] Adopt edge intelligent decision-making algorithms to optimize decisions in real time, such as scheduling adjustments, fault diagnosis, emergency response, etc., to improve system response speed and accuracy.

[0096] Local decision execution:

[0097] Edge computing nodes make decisions such as scheduling, control, and security response based on local computing results, reducing data transmission delays while ensuring real-time performance.

[0098] Management platform: centralized control and manual intervention

[0099] The management platform ensures the flexibility and efficient operation of the system by centrally displaying and monitoring the system status and providing a manual intervention interface.

[0100] Centralized data display:

[0101] The management platform aggregates all sensor data, scheduling plans, fault status and safety warning information, and displays them through a visual interface.

[0102] The platform provides managers with real-time system status feedback, helping them to identify potential problems and make adjustments in a timely manner.

[0103] Manual intervention control:

[0104] Provides a manual intervention control interface to allow managers to make manual adjustments or interventions when the system's automatic scheduling or early warning fails to meet actual needs.

[0105] Adaptive Optimization:

[0106] The management platform uses adaptive optimization algorithms to adjust the overall performance of the system in real time, optimize scheduling algorithms, routing selection strategies, etc., and improve the efficiency of the system.

[0107] The sensors in the perception layer adopt heterogeneous multimodal sensor data fusion technology to realize the joint processing of multiple sensor data, and adopt adaptive weighted fusion algorithm to enable the system to automatically adjust sensor weights and optimize data accuracy under different environmental conditions.

[0108] The network layer uses an AI-based network load prediction model to predict network traffic and bandwidth requirements in real time, and adopts an intelligent routing dynamic optimization algorithm to reduce network congestion and latency, ensuring high-reliability data transmission.

[0109] In one embodiment, the network layer plays a vital role in the rail transit intelligent control system based on the Internet of Things. It is responsible for data transmission, routing selection and optimization, ensuring that the data collected by the perception layer can be transmitted to the edge computing nodes and application layer stably, with low latency and high reliability. To achieve this goal, the network layer adopts a network load prediction model based on artificial intelligence (AI) and an intelligent routing dynamic optimization algorithm. Its working principle mainly includes the following key steps:

[0110] 1.1 Network Load Prediction

[0111] Real-time network traffic monitoring: The network layer first monitors the current network traffic, bandwidth usage, and data packet transmission delay in real time. All data transmitted through wireless communication modules (such as Wi-Fi, 4G / 5G, LoRaWAN, etc.) will be dynamically tracked through this process.

[0112] Artificial intelligence model training: By collecting a large amount of historical network traffic data, machine learning algorithms (such as deep neural networks, long short-term memory networks, and support vector machines) are used to model network traffic and bandwidth requirements.

[0113] Deep Neural Network (DNN): Through the network historical data and real-time data traffic characteristics, a multi-layer neural network model is trained to predict the network load changes in the short term in the future.

[0114] Long Short-Term Memory (LSTM) network: For time series data, the LSTM model can effectively capture the time dependency of network traffic and predict network bandwidth requirements and traffic change trends by memorizing past traffic changes.

[0115] Prediction output: Based on the trained AI model, the system can predict the network traffic and bandwidth requirements in the future in real time. For example, the system predicts the data transmission volume between a station or train in a certain period of time in the future, so as to arrange the appropriate route and bandwidth for the next data traffic.

[0116] 1.2 Intelligent Routing Dynamic Optimization Algorithm

[0117] Route selection: Based on the predicted network load information, the network layer uses an intelligent routing selection algorithm to optimize the data transmission path. The core of the intelligent routing algorithm is to dynamically select the best route based on the real-time network load prediction results.

[0118] Routing selection based on genetic algorithms: Through genetic algorithms, the system performs multi-objective optimization and selects network paths based on factors such as network traffic, bandwidth requirements, latency, network topology, etc. Genetic algorithms iteratively solve the optimal path through operations such as selection, crossover, and mutation.

[0119] Particle Swarm Optimization (PSO): Particle Swarm Optimization (PSO) finds the optimal routing path by simulating the movement of individuals in a group. When selecting a route, PSO can comprehensively consider multiple factors such as network traffic, transmission delay, bandwidth, etc. to ensure that data traffic can be transmitted as smoothly and efficiently as possible.

[0120] Real-time route adjustment: As network traffic changes in real time, the system will constantly monitor the status of the network and dynamically adjust the transmission path. Especially when the network is heavily loaded, the system can dynamically adjust the route selection to avoid network congestion and bandwidth bottlenecks on certain routes. Through such adjustments, the network system can ensure the smooth transmission of data streams and avoid problems such as packet loss, excessive latency or overload.

[0121] Feedback mechanism of routing optimization algorithm: Intelligent routing optimization algorithm has adaptive ability and can optimize according to changes in network status. The system will dynamically select the best path according to the real-time changes in network traffic and the results of load prediction, and respond in time when the network load is too high, such as selecting an alternative path or reducing the transmission bandwidth requirement.

[0122] 1.3 Collaborative working principle;

[0123] Real-time data monitoring and AI prediction synergy: Real-time network traffic and bandwidth demand monitoring combined with AI load prediction model ensures optimal network traffic and bandwidth utilization through dynamically optimized data transmission paths. AI load prediction model provides data support for route optimization, while intelligent route optimization algorithm relies on accurate prediction results to select the optimal path.

[0124] Route optimization and bandwidth management work together: When bandwidth demand peaks, the intelligent route optimization algorithm can not only select low-latency paths, but also optimize the allocation of bandwidth resources. When the network is under high load, the system can automatically adjust bandwidth resources to avoid network bottlenecks.

[0125] The intelligent dispatching module adopts a multi-objective optimization algorithm based on the combination of genetic algorithm and particle swarm optimization algorithm. It performs dispatching calculations based on real-time train position, speed, track condition, traffic flow and system load data, generates train operation plans, and dynamically optimizes the train's operation path, speed, stop time, etc., further improving the response speed and accuracy of train dispatching, reducing delays, energy consumption and congestion.

[0126] It effectively improves the reliability, low latency and efficiency of data transmission. By predicting network traffic and bandwidth requirements in real time and optimizing data transmission paths with intelligent routing algorithms, the system can adaptively adjust in complex network environments, thereby greatly improving the response speed and operating efficiency of the entire rail transit system.

[0127] The fault prediction module adopts a machine learning method based on a combination of deep neural networks and time series models. It accurately predicts equipment failures through a comprehensive analysis of historical data, real-time data, and equipment operating parameters. The algorithm is continuously optimized through reinforcement learning to improve prediction accuracy and reduce the occurrence of failures.

[0128] In one embodiment, the fault prediction module provides accurate fault prediction capabilities in the rail transit intelligent control system by combining deep neural networks (DNN), LSTM time series models and reinforcement learning. It can not only monitor the operating status of the equipment in real time and identify potential fault risks in advance, but also continuously optimize the prediction algorithm through reinforcement learning, thereby improving the prediction accuracy and response speed of the system. This module greatly improves the reliability of the equipment, reduces the occurrence of faults, and optimizes the operation and maintenance costs of the system.

[0129] The safety warning module integrates data from infrared thermal imaging sensors, smoke sensors, vibration sensors, acceleration sensors and environmental gas sensors, and uses deep learning-based abnormal pattern recognition technology. When fire, derailment, abnormal vibration or other safety incidents are detected, an alarm is immediately issued, and multi-level safety response measures are triggered through the automated control system, including automatic parking, evacuation instructions and warning broadcasts.

[0130] In one embodiment, the safety warning module plays a vital role in the rail transit intelligent control system by deeply integrating multiple sensor data and combining deep learning abnormal pattern recognition technology. It can monitor the safety status of equipment, tracks and environment in real time, quickly identify potential security threats, and initiate multi-level safety response measures through automated control systems to ensure the safety of passengers and equipment. This module not only improves the accuracy and response speed of safety warnings, but also significantly reduces the accident rate, operational losses and safety risks, providing strong guarantees for the intelligent and automated management of rail transit systems.

[0131] Edge computing nodes perform real-time calculations on large amounts of sensor data through local data processing and aggregation technology, and make scheduling optimization, fault detection or safety alarm decisions based on the calculation results. The decisions are executed locally by the edge computing nodes to reduce data transmission delays.

[0132] In one embodiment, in the rail transit intelligent control system based on the Internet of Things, edge computing nodes play a vital role, mainly responsible for real-time calculation, local processing and decision execution of a large amount of sensor data. The working principle of edge computing nodes is based on local data processing and aggregation, so that data processing and decision-making can be completed quickly locally, thereby significantly reducing data transmission delays and improving system response speed and reliability;

[0133] 1.1 Data Collection and Transmission

[0134] Edge computing nodes are connected to a large number of sensors and devices, collecting data from various sensors in the rail transit system, including dynamic data of trains (such as speed, acceleration, position, etc.), track status data (such as temperature, pressure, vibration, etc.), environmental data in the carriage (such as temperature, humidity, smoke concentration, etc.), and other key monitoring data. Sensor data is transmitted to edge computing nodes through wireless communication modules (such as Zigbee, NB-I oT, Wi-Fi, etc.).

[0135] Data collection: Edge computing nodes interact with the perception layer through wireless networks and receive data from different sensors in real time.

[0136] Local data aggregation: In the edge computing node, the collected data is first aggregated and preprocessed, such as noise removal, data standardization, data fusion, etc. The aggregation process helps to reduce the redundant data transmitted and ensure data quality.

[0137] 1.2 Data processing and calculation

[0138] Edge computing nodes perform local data processing and calculations to quickly identify potential failures, abnormal conditions, or scheduling optimization needs. The main tasks include:

[0139] Data preprocessing and feature extraction:

[0140] The edge computing node first preprocesses the received raw data (such as denoising, missing value filling, etc.), and then extracts feature information. The feature extraction process helps to extract valuable information from large amounts of data and reduce computational complexity.

[0141] Data aggregation and filtering: Through local aggregation and filtering algorithms (such as moving average method, Kalman filtering, etc.), edge computing nodes can clean up noisy data, remove redundant information, and obtain more accurate sensor data.

[0142] Real-time calculation and analysis:

[0143] Edge computing nodes perform various computing tasks based on real-time data, such as:

[0144] Scheduling Optimization: Optimize train scheduling plans based on real-time train location, speed, track conditions, passenger flow and other information.

[0145] Fault detection: Use sensor data to analyze the operating status of the equipment and detect potential faults or abnormal conditions through machine learning algorithms (such as deep neural networks, support vector machines, etc.).

[0146] Safety warning: By integrating multiple sensor data such as temperature, smoke, vibration, etc., it is possible to determine whether there are safety hazards such as fire, derailment, equipment failure, etc.

[0147] Decision execution and local control:

[0148] After completing data processing, the edge computing node makes decisions based on the analysis results. The decisions include scheduling optimization instructions, fault alarms, or triggering safety warnings.

[0149] Local execution of decisions: Once a decision is made, the edge computing node can directly execute relevant instructions, such as adjusting train operations, triggering fault alarms, or executing emergency responses. This local execution method greatly reduces data transmission delays and improves the system's real-time response capabilities

[0150] The system further includes a management platform, which is used to centrally display all sensor data, scheduling plans, fault status and safety warning information, and provides a manual intervention control interface to allow managers to make necessary adjustments to the system.

[0151] Embodiment 2:

[0152] Please see attached Figure 1 This embodiment provides a rail transit intelligent control method based on the Internet of Things on the basis of the above embodiment, including the following steps:

[0153] Step 1: Use multiple heterogeneous sensors to collect various data of the rail transit system in real time, including train status, track environment, car environment and passenger behavior data, and use adaptive data fusion algorithm for data preprocessing;

[0154] Step 2: The collected data is transmitted to the edge computing node through the wireless communication protocol for preliminary processing. Data preprocessing includes denoising, standardization, anomaly detection, etc., and the transmission path is optimized through the intelligent routing selection algorithm based on deep learning;

[0155] Step 3: Perform fusion analysis on the processed data at the application layer, predict equipment failures based on historical data and real-time data, generate scheduling optimization plans, and issue safety warnings. Fault prediction is performed by combining deep learning with time series analysis algorithms.

[0156] Step 4: Generate the train's dispatch path, speed and stop time based on the data analysis results, and dynamically optimize the dispatch plan through the intelligent dispatch module and send it to the train for execution;

[0157] Step 5: When a potential fault or safety hazard is detected, emergency response measures are immediately triggered, and decisions are locally executed through edge computing nodes to adjust train operations or stop operations and issue safety alerts.

[0158] The method adopts a scheduling optimization algorithm based on a combination of deep neural network and particle swarm optimization algorithm. The method also adopts an equipment fault prediction method that combines convolutional neural network and support vector machine. The method also uses a self-supervised learning algorithm for real-time data analysis and adjusts the scheduling, control and alarm strategies in time according to the prediction results. The algorithm is continuously optimized through reinforcement learning to improve the system prediction accuracy and response speed.

[0159] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The rail transit intelligent control system based on the Internet of Things is characterized by: The system comprises: The perception layer collects various data of the rail transit system in real time by integrating multiple sensors such as visual sensors, thermal imaging sensors, sound sensors, pressure sensors, temperature and humidity sensors, and CO2 sensors. The perception layer uses an adaptive data fusion method and adopts deep learning and reinforcement learning algorithms to jointly optimize and process sensor data to achieve high-precision real-time monitoring, data quality screening, and anomaly detection; The network layer, including the wireless communication module, adopts an intelligent routing selection mechanism based on real-time network load prediction, dynamically selecting the data transmission path according to the real-time network status and load to ensure the stability and low latency of data transmission; The application layer includes an intelligent dispatching module, a fault prediction module and a safety warning module. The intelligent dispatching module dynamically adjusts the train dispatching plan according to real-time data by adopting a multi-objective optimization algorithm. The fault prediction module jointly analyzes historical data and real-time data through a deep neural network and a self-supervised learning algorithm to predict equipment failures and generate warning signals. The safety warning module determines potential safety hazards based on multiple sensor data and triggers automatic response measures. Edge computing nodes are used for real-time computing and analysis of local data, generating scheduling instructions, fault warnings or safety warnings, and reducing data transmission delays and optimizing local decision responses through edge intelligent decision-making algorithms; The management platform is used to centrally display all sensor data, scheduling plans, fault status, and early warning information, and provide a manual intervention control interface to allow managers to make necessary adjustments to the system. The platform further uses an adaptive optimization algorithm to adjust the overall performance of the system.

2. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The sensors of the perception layer adopt heterogeneous multimodal sensor data fusion technology to realize the joint processing of multiple sensor data, and adopt adaptive weighted fusion algorithm to enable the system to automatically adjust sensor weights under different environmental conditions and optimize data accuracy.

3. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The network layer uses an artificial intelligence-based network load prediction model to predict network traffic and bandwidth requirements in real time, and adopts an intelligent routing dynamic optimization algorithm to reduce network congestion and delays, ensuring high-reliability data transmission.

4. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The intelligent scheduling module adopts a multi-objective optimization algorithm based on a combination of genetic algorithm and particle swarm optimization algorithm to perform scheduling calculations based on real-time train position, speed, track condition, traffic flow and system load data, generate train operation plans, and dynamically optimize the train's operation path, speed, stop time, etc., to further improve the response speed and accuracy of train scheduling, reduce delays, energy consumption and congestion.

5. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The fault prediction module adopts a machine learning method based on a combination of deep neural networks and time series models. It accurately predicts equipment failures through a comprehensive analysis of historical data, real-time data, and equipment operating parameters. The algorithm is continuously optimized through reinforcement learning to improve prediction accuracy and reduce the occurrence of failures.

6. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The safety warning module integrates data from infrared thermal imaging sensors, smoke sensors, vibration sensors, acceleration sensors and environmental gas sensors, and adopts abnormal pattern recognition technology based on deep learning. When fire, derailment, abnormal vibration or other safety incidents are detected, an alarm is immediately issued, and multi-level safety response measures are triggered through the automated control system, including automatic parking, evacuation instructions and warning broadcasts.

7. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The edge computing node performs real-time calculations on a large amount of sensor data through local data processing and aggregation technology, and makes scheduling optimization, fault detection or safety alarm decisions based on the calculation results. The decisions are executed locally by the edge computing node to reduce data transmission delays.

8. The rail transit intelligent control system based on the Internet of Things according to claim 1 is characterized in that: The system further includes a management platform, which is used to centrally display all sensor data, scheduling plans, fault status and safety warning information, and provide a manual intervention control interface to allow managers to make necessary adjustments to the system.

9. An intelligent rail transit control method based on the Internet of Things, according to the intelligent rail transit control system based on the Internet of Things according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: Use multiple heterogeneous sensors to collect various data of the rail transit system in real time, including train status, track environment, car environment and passenger behavior data, and use adaptive data fusion algorithm for data preprocessing; Step 2: The collected data is transmitted to the edge computing node for preliminary processing through a wireless communication protocol. The data preprocessing includes denoising, standardization, anomaly detection, etc., and the transmission path is optimized through an intelligent routing selection algorithm based on deep learning; Step 3: Perform fusion analysis on the processed data at the application layer, predict equipment failures based on historical data and real-time data, generate scheduling optimization plans, and issue safety warnings. The failure prediction is performed by combining deep learning with time series analysis algorithms. Step 4: Generate the train's dispatch path, speed and stop time based on the data analysis results, and dynamically optimize the dispatch plan through the intelligent dispatch module and send it to the train for execution; Step 5: When a potential fault or safety hazard is detected, emergency response measures are immediately triggered, and decisions are locally executed through edge computing nodes to adjust train operations or stop operations and issue safety alerts.

10. The rail transit intelligent control method based on the Internet of Things according to claim 9 is characterized in that: The method adopts a scheduling optimization algorithm based on a combination of a deep neural network and a particle swarm optimization algorithm. The method also adopts an equipment fault prediction method that combines a convolutional neural network and a support vector machine. The method also uses a self-supervised learning algorithm to perform real-time data analysis and timely adjusts scheduling, control and alarm strategies according to prediction results. The algorithm is continuously optimized through reinforcement learning to improve system prediction accuracy and response speed.

Images