Urban rail transit intelligent operation management method and device

Abstract

The invention relates to the technical field of artificial intelligence, and particularly provides an urban rail transit intelligent operation management method and device, and the method comprises the following steps: S1, constructing a multi-source heterogeneous data integration platform; s2, constructing a knowledge extraction and understanding module; s3, the image recognition and analysis module performs image feature extraction, image analysis and processing and thematic chart generation; s4, combining a digital twinning technology and a physical simulation engine to realize combined calling of the rail transit professional model; s5, interactive document office and knowledge management; s6, a standardized interface protocol is developed, so that the large rail transit model can be directly connected with various terminal devices and facility control systems, and remote monitoring and control are achieved; and S7, constructing a passenger-oriented service platform, and providing real-time information query, navigation guidance and ticket booking. Compared with the prior art, the operation management efficiency of the urban rail transit can be greatly improved, the safety is enhanced, and the travel experience of passengers is optimized.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and specifically provides a method and device for intelligent operation management of urban rail transit. Background Art

[0002] With the acceleration of urbanization, urban rail transit, as an important public transportation tool, has played an important role in alleviating urban traffic congestion and improving travel efficiency.

[0003] However, traditional urban rail transit systems still face many challenges in operational management and passenger services, including but not limited to: inaccurate passenger flow forecasts: traditional passenger flow forecasting methods often rely on historical data and experience, and are difficult to cope with emergencies and complex changes; low operational efficiency: the efficiency of train scheduling, maintenance, fault diagnosis, etc. needs to be improved; poor passenger experience: passenger services in ticket purchase, transfer, navigation, etc. are not intelligent enough, resulting in a poor experience; insufficient security management level: traditional security measures are difficult to cope with increasingly complex public safety threats. Summary of the invention

[0004] The present invention aims at the above-mentioned deficiencies of the prior art and provides a highly practical urban rail transit intelligent operation and management method.

[0005] A further technical task of the present invention is to provide an intelligent operation and management device for urban rail transit that is reasonably designed, safe and applicable.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] An urban rail transit intelligent operation management method comprises the following steps:

[0008] S1. Build a multi-source heterogeneous data integration platform;

[0009] S2. Build a knowledge extraction and understanding module based on the big model to carry out customized language model training, knowledge graph construction, task decomposition and execution, and parameter fine-tuning;

[0010] S3, image recognition and analysis module performs image feature extraction, image analysis and processing, and thematic chart generation;

[0011] S4. Combine digital twin technology and physical simulation engine to realize the combined call of rail transit professional models;

[0012] S5, interactive document office and knowledge management;

[0013] S6. Develop standardized interface protocols to enable large rail transit models to directly connect with various terminal equipment and facility control systems to achieve remote monitoring and control;

[0014] S7. Build a passenger-oriented service platform to provide real-time information query, navigation guidance and ticket booking.

[0015] Furthermore, in step S1, it includes:

[0016] (1) Data collection and transmission: Various sensors are deployed. The sensors transmit data to the data center in real time through wireless or wired means. The real-time data of the train's position, speed, and acceleration are collected through on-board sensors and train control systems, and transmitted to the data center through a dedicated communication network. The existing ticketing system is integrated to collect passenger entry and exit data and ticket purchase records, and access meteorological data, traffic data, and social media data public databases to provide more comprehensive data support.

[0017] (2) Data preprocessing: removing invalid data, duplicate data, and noisy data, converting data from different sources into a unified format, normalizing the data, and verifying the integrity and consistency of the data through verification algorithms;

[0018] (3) Use Hadoop HDFS and Ceph distributed storage systems, establish data indexing mechanisms, perform regular data backups, and use encryption technology and access control measures;

[0019] (4) Dynamic update and real-time monitoring: Apache Kafka and Apache Flink real-time stream processing frameworks are used to process and analyze sensor data in real time. Through the data synchronization mechanism, real-time data updates are ensured, and a real-time data visualization interface is provided.

[0020] Furthermore, in step S1, in the customized training of the language model, a pre-trained language model suitable for the field of urban rail transit is selected, a large amount of rail transit literature, reports and technical specifications are used as training corpus, the pre-trained model is fine-tuned, and the professionalism of the model is further enhanced by combining professional terminology, business rules and expert experience in the field of rail transit;

[0021] The knowledge graph is constructed by extracting key entities from text through named entity recognition technology, using a relationship extraction algorithm to identify the relationships between entities, and storing the extracted entities and relationships in a graph database to form a knowledge graph in the field of rail transit.

[0022] Furthermore, in the task decomposition and execution, complex operation management tasks are decomposed into multiple subtasks, and the subtasks are understood and executed using a large model to generate corresponding solutions. Through a multi-scale similarity algorithm, historical operation data and preset deduction processes are matched to assist in formulating accurate and effective operation plans.

[0023] In the parameter fine-tuning, the model is regularly updated comprehensively, and the LoRA parameter efficient fine-tuning method is adopted.

[0024] Furthermore, in step S3, when image feature extraction is performed, abnormal behaviors on the platform and in the carriage are detected through a behavior recognition algorithm, equipment failures are automatically detected through image recognition technology, and passenger flow distribution on the platform and in the carriage is monitored in real time through image analysis to generate a passenger flow heat map;

[0025] During image analysis and processing, the behavior recognition algorithm is used to detect abnormal behaviors on the platform and in the carriages. The image recognition technology is used to automatically detect equipment failures. The image analysis is used to monitor the passenger flow distribution on the platform and in the carriages in real time and generate a passenger flow heat map.

[0026] When generating thematic charts, they are dynamically generated according to the themes set by the user, providing interactive chart displays and automatically generating image analysis reports.

[0027] Furthermore, in step S4, it includes:

[0028] (1) Digital twin technology: Establish a digital twin model of the urban rail transit system. Use virtual simulation technology to simulate actual operation conditions, synchronize real-time data to the digital twin model, and conduct multi-dimensional data analysis through the digital twin model.

[0029] (2) Physical simulation engine: Use the train dispatching model to simulate train operation and optimize the train operation plan. Use the passenger flow prediction model to predict the passenger flow in the future and assist in formulating a reasonable operation plan. Use the energy management model to optimize energy consumption.

[0030] (3) Multi-scale similarity matching: Through the multi-scale similarity algorithm, historical operation data is matched to find similar operation scenarios. Based on historical data and preset deduction processes, an effective operation plan is generated. Combined with real-time data, the operation plan is dynamically adjusted.

[0031] (4) Operation plan generation: Generate train dispatching plans based on simulation and prediction results, optimize train operation plans, automatically generate emergency plans, and automatically generate operation reports.

[0032] Furthermore, in step S5, it includes:

[0033] (1) Automatically generate the first draft of operation reports and data analysis documents based on the user's requirements and specified templates. Through natural language processing technology, it supports users to edit and modify documents through dialogue, providing document management functions;

[0034] (2) Knowledge base management: establishing a rich public knowledge base, allowing users to create private knowledge bases, storing proprietary knowledge and experience within the enterprise, and creating a knowledge retrieval mechanism;

[0035] (3) Collaborative office, supports collaborative document editing by multiple people, provides version control functions, and sets permissions for different users.

[0036] Furthermore, in step S6, it includes:

[0037] (1) Standardized interface protocols, defining unified interface specifications and adopting standardized communication protocols;

[0038] (2) Through sensors and cameras, the device status is monitored in real time, supporting remote control of the device, integrating fault diagnosis algorithms, automatically detecting potential problems and alerting relevant personnel;

[0039] (3) Security and privacy protection: encrypt the transmitted data, set up a strict access control mechanism, only authorized users can access and control the device, and record all operation logs.

[0040] Furthermore, in step S7, it includes:

[0041] (1) Real-time information query, providing train timetable query, station information query and fare information query;

[0042] (2) Navigation guidance: providing route planning services to help passengers find the shortest and fastest route. It provides real-time navigation guidance and detailed transfer guidance through GPS positioning and map data.

[0043] (3) Ticket reservation, online ticket purchase services, provision of electronic ticket services and ticket refund and change services;

[0044] (4) Natural language processing: through natural language processing technology, passengers can query information and book tickets by voice, and passengers can query information and book tickets by text in multiple languages;

[0045] (5) Personalized services: By analyzing passengers’ historical behavior data, we can understand their travel habits and preferences. Based on their travel habits and preferences, we can provide personalized travel suggestions and services, provide user feedback channels, collect passengers’ opinions and suggestions, and continuously optimize services.

[0046] An intelligent operation and management device for urban rail transit, comprising: at least one memory and at least one processor;

[0047] The at least one memory is used to store a machine-readable program;

[0048] The at least one processor is used to call the machine-readable program to execute a method for intelligent operation and management of urban rail transit.

[0049] Compared with the prior art, the urban rail transit intelligent operation management method and device of the present invention have the following outstanding beneficial effects:

[0050] Through the application of the above measures, the present invention is expected to significantly improve many problems existing in the existing urban rail transit management system, such as serious information island phenomenon and slow response. More importantly, it can also promote interdisciplinary exchanges and cooperation and promote the development of related theories and practices. Specifically, the system can play an important role in the following aspects:

[0051] Improve the operational efficiency of urban rail transit systems and reduce delays and outages;

[0052] Enhance system security, timely discover and handle potential safety hazards through real-time monitoring and intelligent analysis;

[0053] Optimize passengers' travel experience and provide convenient and efficient services;

[0054] Reduce operating costs and improve resource utilization;

[0055] Provide solid technical support for the intelligent transformation of urban rail transit. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Attached Figure 1 The invention is a flowchart of an intelligent operation and management method for urban rail transit. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.

[0059] A best embodiment is given below:

[0060] like Figure 1As shown, a method for intelligent operation and management of urban rail transit in this embodiment has the following steps:

[0061] S1. Build a multi-source heterogeneous data integration platform;

[0062] Establish a unified portal that can receive relevant information from different sensor networks (such as cameras, temperature sensors, vibration sensors, etc.), train operation data, ticketing systems, and other public databases, and clean and convert them through preprocessing algorithms to ensure the quality of data used for subsequent analysis. The platform should be able to dynamically update information in the urban rail transit network, including data on key objects such as train location, station status, passenger flow distribution, etc., while integrating high-resolution video surveillance streams and other sensor data.

[0063] (1) Data collection and transmission;

[0064] Sensor network: Deploy various sensors, such as cameras, temperature sensors, vibration sensors, etc., to cover the entire urban rail transit network. Sensors transmit data to the data center in real time via wireless or wired means.

[0065] Train operation data: Real-time data such as the train’s position, speed, acceleration, etc. are collected through on-board sensors and train control systems, and transmitted to the data center through a dedicated communication network.

[0066] Ticketing system: Integrate the existing ticketing system to collect passenger entry and exit data, ticket purchase records and other information.

[0067] Public databases: Access public databases such as meteorological data, traffic data, and social media data to provide more comprehensive data support.

[0068] (2) Data preprocessing;

[0069] Data cleaning: remove invalid data, duplicate data and noisy data to ensure data quality.

[0070] Format conversion: Convert data from different sources into a unified format for easy subsequent processing.

[0071] Data standardization: Normalize the data to ensure that data from different sources are comparable on the same scale.

[0072] Data verification: Verify the integrity and consistency of data through verification algorithms to prevent data loss or damage.

[0073] (3) Data storage and management;

[0074] Distributed storage: Distributed storage systems such as Hadoop HDFS and Ceph are used to ensure high availability and scalability of data.

[0075] Data indexing: Establish an efficient data indexing mechanism to improve the speed of data query and retrieval.

[0076] Data backup: Perform data backup regularly to prevent data loss and ensure system reliability.

[0077] Data security: Encryption technology, access control and other means are used to ensure data security and privacy.

[0078] (4) Dynamic update and real-time monitoring;

[0079] Real-time data stream processing: Use real-time stream processing frameworks such as Apache Kafka and Apache Flink to process and analyze sensor data in real time.

[0080] Dynamic update: Through the data synchronization mechanism, ensure the real-time update of data and keep the data up to date.

[0081] Data visualization: Provides a real-time data visualization interface to display key information such as train location, station status, passenger flow distribution, etc., which is convenient for operators to monitor.

[0082] S2. Build a knowledge extraction and understanding module based on the big model to carry out customized language model training, knowledge graph construction, task decomposition and execution, and parameter fine-tuning;

[0083] We use language models trained on large-scale texts and combine them with professional terminology, business rules, and expert experience in the field of urban rail transit for customized training. We use knowledge graph technology to enhance the model's understanding of rail transit expertise, enabling it to decompose complex operational management tasks, such as predicting train scheduling plans based on real-time passenger flow. In addition, we use efficient parameter fine-tuning methods such as full update or LoRA to further enhance the model's understanding and task execution capabilities for specific rail transit business scenarios.

[0084] (1) Customized language model training;

[0085] Pre-trained model selection: Select pre-trained language models suitable for the field of urban rail transit, such as BERT, GPT-3 / 4, etc.

[0086] Customized fine-tuning: Use a large amount of rail transit literature, reports, and technical specifications as training corpus to fine-tune the pre-trained model.

[0087] Domain knowledge integration: Combine professional terminology, business rules and expert experience in the field of rail transit to further enhance the professionalism of the model.

[0088] (2) Knowledge graph construction;

[0089] Entity recognition: Through named entity recognition technology, key entities such as stations, trains, equipment, etc. are extracted from the text.

[0090] Relationship extraction: Use relationship extraction algorithms to identify relationships between entities, such as the relationship between trains and stations, the relationship between equipment and faults, etc.

[0091] Knowledge graph storage: The extracted entities and relationships are stored in a graph database to form a knowledge graph in the field of rail transit.

[0092] (3) Task decomposition and execution;

[0093] Task decomposition: Decompose complex operation management tasks into multiple subtasks, such as passenger flow forecasting, train scheduling, equipment maintenance, etc.

[0094] Task execution: Use the big model to understand and execute subtasks and generate corresponding solutions.

[0095] Multi-scale similarity matching: Through the multi-scale similarity algorithm, historical operation data and preset deduction processes are matched to assist in formulating accurate and effective operation plans.

[0096] (4) Efficient parameter fine-tuning;

[0097] Full update: Regularly update the model in full to ensure its currency and accuracy.

[0098] LoRA fine-tuning: Use LoRA (Low-Rank Adaptation) and other efficient parameter fine-tuning methods to reduce training time and computing resource consumption, and improve the adaptability and flexibility of the model

[0099] S3, image recognition and analysis module performs image feature extraction, image analysis and processing, and thematic chart generation;

[0100] It integrates advanced computer vision technology to automatically identify and classify features in high-resolution video surveillance images, such as identifying platform crowding, detecting abnormal behavior, identifying equipment failures, etc. This module also supports users to set themes and dynamically generate thematic charts of key information such as passenger flow distribution and train operation status, providing users with intuitive and accurate information display.

[0101] (1) Image feature extraction;

[0102] Convolutional Neural Network (CNN): Convolutional neural network is used to extract image features and identify key elements in the image, such as people, equipment, obstacles, etc.

[0103] Target detection: Use target detection algorithms such as Faster R-CNN and YOLO to automatically detect specific targets in images, such as abnormal behavior, equipment failure, etc.

[0104] Image classification: Images are classified by image classification algorithms, such as platform congestion level, train operation status, etc.

[0105] (2) Image analysis and processing;

[0106] Abnormal behavior detection: Through behavior recognition algorithms, abnormal behaviors on platforms and in carriages, such as passengers falling or fighting, can be detected.

[0107] Equipment fault identification: Through image recognition technology, equipment faults such as track cracks, signal light failures, etc. can be automatically detected.

[0108] Passenger flow analysis: Through image analysis, the passenger flow distribution on the platform and in the carriage is monitored in real time to generate a passenger flow heat map.

[0109] (3) Generation of thematic charts;

[0110] Data visualization: Dynamically generate thematic charts of key information such as passenger flow distribution and train operation status based on the theme set by the user.

[0111] Interactive charts: Provide interactive chart displays, where users can view detailed information and historical data by clicking, dragging, etc.

[0112] Report generation: Automatically generate image analysis reports, including abnormal behavior detection results, equipment failure conditions, passenger flow distribution, etc.

[0113] S4. Combine digital twin technology and physical simulation engine to realize the combined call of rail transit professional models;

[0114] Combining digital twin technology and physical simulation engine, it realizes the combined call of rail transit professional models, such as train scheduling model, passenger flow prediction model, energy management model, etc. Through these models, the system can perform minute-level train operation simulation, passenger flow prediction, energy consumption optimization, etc. On this basis, with the help of multi-scale similarity algorithm matching technology driven by large models, it realizes the similarity matching of operation scenarios based on historical operation data and preset deduction process, so as to assist in formulating more accurate and effective operation plans.

[0115] (1) Digital twin technology;

[0116] Virtual simulation: Establish a digital twin model of the urban rail transit system and simulate actual operation conditions through virtual simulation technology.

[0117] Real-time data synchronization: Synchronize real-time data to the digital twin model to ensure that the virtual model is consistent with the actual situation.

[0118] Multi-dimensional analysis: Through the digital twin model, multi-dimensional data analysis can be carried out, such as train operation efficiency, energy consumption, passenger flow distribution, etc.

[0119] (2) Physical simulation engine;

[0120] Train scheduling model: Use the train scheduling model to simulate train operation and optimize train operation plans.

[0121] Passenger flow prediction model: The passenger flow prediction model can be used to predict passenger flow in the future and assist in formulating reasonable operation plans.

[0122] Energy management model: Through the energy management model, optimize energy consumption and improve energy utilization efficiency.

[0123] (3) Multi-scale similarity matching;

[0124] Historical data matching: Use a multi-scale similarity algorithm to match historical operation data and find similar operation scenarios.

[0125] Preset deduction process: Generate accurate and effective operation plans based on historical data and preset deduction process

[0126] Real-time optimization: Combine real-time data to dynamically adjust operation plans to ensure operational efficiency and flexibility.

[0127] (4) Operation plan generation;

[0128] Scheduling plan recommendation: Generate train scheduling plans based on simulation and prediction results to optimize train operation plans.

[0129] Emergency Plan: Automatically generate emergency plans to deal with emergencies such as equipment failure, bad weather, etc.

[0130] Operation report: Automatically generate operation reports, including train operation status, passenger flow distribution, energy consumption, etc.

[0131] S5, interactive document office and knowledge management;

[0132] It provides an intelligent document preparation assistant that can automatically generate operation reports, write the first draft of data analysis documents, summarize complex materials, etc. according to the needs and specified templates proposed by users. This function not only supports online reading or downloading of generated documents, but also allows collaborative modification and improvement of materials through dialogue. At the same time, the system has built-in rich public knowledge base and private knowledge base retrieval mechanism to ensure that the information source is reliable and comprehensive.

[0133] (1) Intelligent document preparation assistant;

[0134] Automatic document generation: Automatically generate operational reports, data analysis document drafts, etc. based on user requirements and specified templates.

[0135] Document editing and modification: Through natural language processing technology, users are supported to edit and modify documents through conversation.

[0136] Document management: Provides document management functions and supports online reading, downloading, and sharing of documents.

[0137] (2) Knowledge base management;

[0138] Public knowledge base: Establish a rich public knowledge base that includes professional knowledge, technical specifications, case analysis, etc. in the field of rail transit.

[0139] Private knowledge base: supports users to create private knowledge base to store proprietary knowledge and experience within the enterprise.

[0140] Knowledge retrieval: Provides an efficient knowledge retrieval mechanism that supports multiple retrieval methods such as keyword search and full-text search.

[0141] (3) Collaborative office work;

[0142] Multi-person collaboration: Supports multiple people to edit documents together to improve work efficiency.

[0143] Version control: Provides version control function to record every modification of the document for easy backtracking and recovery.

[0144] Permission management: Set permissions for different users to ensure the security and confidentiality of documents.

[0145] S6. Develop standardized interface protocols to enable large rail transit models to directly connect with various terminal equipment and facility control systems to achieve remote monitoring and control;

[0146] Develop standardized interface protocols so that large rail transit models can be directly connected to various terminal devices (such as cameras, sensors) and facility control systems to achieve remote monitoring and control. This will help improve all-day and all-weather monitoring and intelligent service capabilities, and also facilitate rapid response to emergencies in emergency situations.

[0147] (1) Standardized interface protocols;

[0148] Interface definition: Define unified interface specifications to ensure that devices produced by different manufacturers can be connected to the system without obstacles.

[0149] Communication protocol: Use standardized communication protocols, such as HTTP / HTTPS, MQTT, etc., to ensure the security and reliability of data transmission.

[0150] (2) Remote monitoring and control;

[0151] Equipment status monitoring: Use sensors and cameras to monitor the status of equipment in real time, such as temperature, humidity, vibration, etc.

[0152] Device control: Support remote control of devices, such as switching devices on and off, adjusting device parameters, etc.

[0153] Fault diagnosis and early warning: Integrated fault diagnosis algorithms automatically detect potential problems and issue alerts to relevant personnel.

[0154] (3) Security and privacy protection;

[0155] Data encryption: Encrypt the transmitted data to prevent it from being stolen or tampered with.

[0156] Access control: Set up a strict access control mechanism so that only authorized users can access and control the device.

[0157] Log audit: record all operation logs to facilitate post-audit and tracking.

[0158] S7. Build a service platform for passengers to provide real-time information query, navigation guidance and ticket booking;

[0159] Build a service platform for passengers, providing real-time information query, navigation guidance, ticket booking and other functions. Through natural language processing technology, voice and text interaction can be realized, so that passengers can obtain the required information in a more natural way. In addition, the system can also provide personalized travel suggestions and services for passengers based on their historical behavior data.

[0160] (1) Real-time information query

[0161] Train timetable: Provide train timetable query, including departure time, arrival time, etc.

[0162] Station information: Provides station information query, including station location, transfer information, etc.

[0163] Fare information: Provides fare information enquiry, including one-way fares, discounted fares, etc.

[0164] (2) Navigation guidance;

[0165] Route planning: Provide route planning services to help passengers find the shortest and fastest route.

[0166] Real-time navigation: Provide real-time navigation guidance through GPS positioning and map data to help passengers reach their destination smoothly.

[0167] Transfer Guide: Provide detailed transfer guide, including transfer stations, transfer routes, etc.

[0168] (3)Ticket booking;

[0169] Online ticket purchase: Supports online ticket purchase service, passengers can purchase tickets through mobile APP or website.

[0170] Electronic ticket: Electronic ticket service is provided, and passengers can use QR code or electronic ticket to enter the station and board the train.

[0171] Refund and Change Tickets: Supports ticket refund and change services to facilitate passengers to change their itineraries.

[0172] (4) Natural language processing;

[0173] Voice interaction: Through natural language processing technology, passengers can query information and book tickets through voice.

[0174] Text interaction: Supports passengers to query information and book tickets via text, providing a natural and smooth interactive experience.

[0175] Multi-language support: Support multiple languages ​​to meet the needs of different passengers.

[0176] (5) Personalized services;

[0177] Historical behavior analysis: Understand passengers’ travel habits and preferences by analyzing their historical behavior data.

[0178] Personalized recommendations: Provide passengers with personalized travel suggestions and services based on their travel habits and preferences.

[0179] User feedback: Provide user feedback channels to collect passengers’ opinions and suggestions and continuously optimize services.

[0180] Based on the above method, an urban rail transit intelligent operation management device in this embodiment includes: at least one memory and at least one processor;

[0181] The at least one memory is used to store a machine-readable program;

[0182] The at least one processor is used to call the machine-readable program to execute a method for intelligent operation and management of urban rail transit.

[0183] The above-mentioned specific implementations are only specific cases of the present invention. The patent protection scope of the present invention includes but is not limited to the above-mentioned specific implementations. Any technical solutions that conform to the above-mentioned specific implementations of the present invention and any appropriate changes or substitutions made by ordinary technicians in the relevant technical field shall fall within the patent protection scope of the present invention.

[0184] 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. A method for intelligent operation and management of urban rail transit, characterized in that: The steps are as follows: S1. Build a multi-source heterogeneous data integration platform; S2. Build a knowledge extraction and understanding module based on the big model to carry out customized language model training, knowledge graph construction, task decomposition and execution, and parameter fine-tuning; S3, image recognition and analysis module performs image feature extraction, image analysis and processing, and thematic chart generation; S4. Combine digital twin technology and physical simulation engine to realize the combined call of rail transit professional models; S5, interactive document office and knowledge management; S6. Develop standardized interface protocols to enable large rail transit models to directly connect with various terminal equipment and facility control systems to achieve remote monitoring and control; S7. Build a passenger-oriented service platform to provide real-time information query, navigation guidance and ticket booking.

2. The urban rail transit intelligent operation management method according to claim 1, characterized in that: In step S1, it includes: (1) Data collection and transmission: Various sensors are deployed. The sensors transmit data to the data center in real time through wireless or wired means. The real-time data of the train's position, speed, and acceleration are collected through on-board sensors and train control systems, and transmitted to the data center through a dedicated communication network. The existing ticketing system is integrated to collect passenger entry and exit data and ticket purchase records, and access meteorological data, traffic data, and social media data public databases to provide more comprehensive data support. (2) Data preprocessing: removing invalid data, duplicate data, and noisy data, converting data from different sources into a unified format, normalizing the data, and verifying the integrity and consistency of the data through verification algorithms; (3) Use Hadoop HDFS and Ceph distributed storage systems, establish data indexing mechanisms, perform regular data backups, and use encryption technology and access control measures; (4) Dynamic update and real-time monitoring: Apache Kafka and Apache Flink real-time stream processing frameworks are used to process and analyze sensor data in real time. Through the data synchronization mechanism, real-time data updates are ensured, and a real-time data visualization interface is provided.

3. The urban rail transit intelligent operation management method according to claim 2 is characterized in that: In step S1, in the customized training of the language model, a pre-trained language model suitable for the field of urban rail transit is selected, a large amount of rail transit literature, reports and technical specifications are used as training corpus, the pre-trained model is fine-tuned, and the professionalism of the model is further enhanced by combining professional terminology, business rules and expert experience in the field of rail transit; The knowledge graph is constructed by extracting key entities from text through named entity recognition technology, using a relationship extraction algorithm to identify the relationships between entities, and storing the extracted entities and relationships in a graph database to form a knowledge graph in the field of rail transit.

4. The urban rail transit intelligent operation management method according to claim 3 is characterized in that: In the task decomposition and execution, complex operation management tasks are decomposed into multiple subtasks, and the subtasks are understood and executed using a large model to generate corresponding solutions. Through a multi-scale similarity algorithm, historical operation data and preset deduction processes are matched to assist in formulating accurate and effective operation plans; In the parameter fine-tuning, the model is regularly updated comprehensively, and the LoRA parameter efficient fine-tuning method is adopted.

5. The urban rail transit intelligent operation management method according to claim 4 is characterized in that: In step S3, when extracting image features, the behavior recognition algorithm is used to detect abnormal behaviors on the platform and in the carriage, and the image recognition technology is used to automatically detect equipment failures. The image analysis is used to monitor the passenger flow distribution on the platform and in the carriage in real time, and a passenger flow heat map is generated; During image analysis and processing, the behavior recognition algorithm is used to detect abnormal behaviors on the platform and in the carriages. The image recognition technology is used to automatically detect equipment failures. The image analysis is used to monitor the passenger flow distribution on the platform and in the carriages in real time and generate a passenger flow heat map. When generating thematic charts, they are dynamically generated according to the themes set by the user, providing interactive chart displays and automatically generating image analysis reports.

6. The urban rail transit intelligent operation management method according to claim 5, characterized in that: In step S4, it includes: (1) Digital twin technology: Establish a digital twin model of the urban rail transit system. Use virtual simulation technology to simulate actual operation conditions, synchronize real-time data to the digital twin model, and conduct multi-dimensional data analysis through the digital twin model. (2) Physical simulation engine: Use the train dispatching model to simulate train operation and optimize the train operation plan. Use the passenger flow prediction model to predict the passenger flow in the future and assist in formulating a reasonable operation plan. Use the energy management model to optimize energy consumption. (3) Multi-scale similarity matching: Through the multi-scale similarity algorithm, historical operation data is matched to find similar operation scenarios. Based on historical data and preset deduction processes, an effective operation plan is generated. Combined with real-time data, the operation plan is dynamically adjusted. (4) Operation plan generation: Generate train dispatching plans based on simulation and prediction results, optimize train operation plans, automatically generate emergency plans, and automatically generate operation reports.

7. The urban rail transit intelligent operation management method according to claim 6 is characterized in that: In step S5, it includes: (1) Automatically generate the first draft of operation reports and data analysis documents based on the user's requirements and specified templates. Through natural language processing technology, it supports users to edit and modify documents through dialogue, providing document management functions; (2) Knowledge base management: establishing a rich public knowledge base, allowing users to create private knowledge bases, storing proprietary knowledge and experience within the enterprise, and creating a knowledge retrieval mechanism; (3) Collaborative office, supports collaborative document editing by multiple people, provides version control functions, and sets permissions for different users.

8. The urban rail transit intelligent operation management method according to claim 7 is characterized in that: In step S6, it includes: (1) Standardized interface protocols, defining unified interface specifications and adopting standardized communication protocols; (2) Through sensors and cameras, the device status is monitored in real time, supporting remote control of the device, integrating fault diagnosis algorithms, automatically detecting potential problems and alerting relevant personnel; (3) Security and privacy protection: encrypt the transmitted data, set up a strict access control mechanism, only authorized users can access and control the device, and record all operation logs.

9. The urban rail transit intelligent operation management method according to claim 8, characterized in that: In step S7, it includes: (1) Real-time information query, providing train timetable query, station information query and fare information query; (2) Navigation guidance: providing route planning services to help passengers find the shortest and fastest route. It provides real-time navigation guidance and detailed transfer guidance through GPS positioning and map data. (3) Ticket reservation, online ticket purchase services, provision of electronic ticket services and ticket refund and change services; (4) Natural language processing: through natural language processing technology, passengers can query information and book tickets by voice, and passengers can query information and book tickets by text in multiple languages; (5) Personalized services: By analyzing passengers’ historical behavior data, we can understand their travel habits and preferences. Based on their travel habits and preferences, we can provide personalized travel suggestions and services, provide user feedback channels, collect passengers’ opinions and suggestions, and continuously optimize services.

10. An intelligent operation and management device for urban rail transit, characterized in that: include: at least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to call the machine-readable program to execute the method according to any one of claims 1 to 9.

Images