This project belongs to the field of rail transit and involves the core technologies for the service safety and intelligent maintenance of key subsystems of urban rail trains—the braking system. Before 2010, research on urban rail train braking systems mainly focused on solving the issue of localization of core technologies in the braking system, while the maintenance and upkeep of the braking system had always followed a regular preventive maintenance strategy. With the gradual completion of the domestication process of the braking system and the rapid development of intelligent technologies such as the Internet of Things and information technology, the research focus of urban rail train braking systems has gradually shifted from localization to the field of intelligent maintenance. The project is funded by the National 863 Program "Research and Development of Robust Detection, Filtering and Networked On-Board Fault Diagnosis System for Braking Systems" (Project Number 2011AA110503-3). The paper for the first time explains the failure behavior law of the braking system as a complex 'machine-electrical-air' coupled system. The system solves the core technical problems of software and hardware in on-the-way fault monitoring and diagnosis, identification and early warning of the braking system under operational conditions. Funded by the 'Urban Rail Transit Train Braking System Service Safety Research' project of the 12th Five-Year National Science and Technology Support Plan (No. 2015BAG19B01-05), it proposes a thought for evaluating the service performance consistency of the braking system and is used for the safety evaluation of the braking systems of urban rail transit trains on 10 domestic subway lines. The main technical innovations include: (1) Based on the physical model driving method, a fault state matrix space under the coupling action of 'machine-electrical-air' in the braking system was constructed. The mapping relationship between fault indices and measured signals was realized, completing the fault characterization of the braking system. (2) The 'first quadrant' law of fault migration change in the braking system was discovered, and for the first time, a fault observable and distinguishable sensor configuration and optimization method based on SDG was proposed. The application of this method significantly improved the number of observable and distinguishable faults in urban rail transit train braking systems. (3) A 'three characteristics' analysis method for evaluating the evolution of service performance of the braking system was proposed, which is used to evaluate the impact of braking efficiency on the safety of rail transit systems. Completed the safety impact assessment of braking service for 10 domestic subway lines, totaling over 2000km. (4) Developed a fault identification and diagnosis software and hardware system for brake systems based on sensor network mode suitable for in-transit monitoring functions, solving the problem of early warning for holographic faults in brake systems, and for the first time realized the monitoring, early warning, and evaluation of air source humidity information for brake systems. The key technologies of the project have independent intellectual property rights, with 5 invention patents granted, 5 software copyright registrations, more than 20 papers published, including 4 SCI-indexed and 13 EI-indexed papers. Completed 4 international academic invited reports, trained 3 doctoral students, 12 master's students, and completed 3 project acceptance inspections. According to the review, the overall technology of the project has reached a leading domestic level. The in-transit fault monitoring and early warning system for brake systems developed in this project consists of distributed sensing monitoring units such as humidity. Data Acquisition and Storage Unit, suitable for onboard network transmission units, 27 types of braking fault online recognition algorithms, etc. The system is scalable and can implement real-time monitoring functions on the line. Fault information can be sent to train display terminals and ground terminals as needed, greatly improving the operational maintenance level of the braking system. The project has been successfully applied by CRRC Zhuzhou Electric Locomotive Co., Ltd. and Guangzhou Metro Group Co., Ltd., with an annual increase in output value of 7.2 million yuan and an annual savings in maintenance costs of 3.42 million yuan. The project's proposed 'three characteristics' analysis method for service effectiveness has completed the evaluation of braking indicators for rail transit safety operations in 10 domestic subway lines, with a cumulative test mileage exceeding 2,000 kilometers. The project's first-time implementation of compressed air humidity monitoring technology for the braking system has filled a domestic gap. It has now been included as one of the common points adopted in urban rail train bidding documents, promoting the advancement and development of industry technology, with significant economic and social benefits.