Tingyu Xin

6G will provide a new opportunity to integrate Artificial Intelligence (AI) and Machine Learning (ML) into wireless communication. The current 5G Core (5GC) architecture possesses limitations that restrict its ability to effectively deploy AI functions, including the lack of a standardized, unified access layer for network data, limited service exposure models that cannot fully support AI feedback loops, and tight coupling between logic and signaling that inhibits the insertion of AI-based decision-making components. This paper explores approaches to address the 5GC shortcomings in the future 6G Core (6GC) network. Recommendations for architectural evolution and standardization are discussed to facilitate scalable, efficient, and adaptable AI-enabled 6G networks.

Railway pantographs are used around the world for collecting electrical energy to power railway vehicles from the overhead catenary. Faults in the pantograph system degrade the quality of the contact between the pantograph and catenary and reduce the reliability of railway operations. To maintain the pantographs in a good working condition, regular inspection tasks are carried out at rolling stock depots. The current pantograph inspections, in general, are only effective for the detection of major faults, providing limited incipient fault detection or fault diagnosis capabilities. Condition monitoring of pantographs has the potential to improve pantograph performance and reduce maintenance costs. As a first step in the realisation of practical pantograph condition monitoring, a laboratory-based pantograph test rig has been developed to gain an understanding of pantograph dynamic behaviours, particularly when incipient faults are present. In the first work of this kind, dynamic response data have been acquired from a number of pantographs that have allowed fault detection and diagnosis algorithms to be developed and verified. Three tests have been developed: (i) a hysteresis test that uses different excitation speeds, (ii) a frequency response test that uses different excitation frequencies, and (iii) a novel changing gradient test. Verification tests indicate that the hysteresis test is effective in detecting and diagnosing pneumatic actuator and elbow joint faults. The frequency response test is able to monitor the overall degradation in the pantograph. The changing gradient test provides fault detection and diagnosis in the pantograph head suspension and pneumatic actuator. The test rig and fault detection and diagnosis algorithms are now being developed into a depot-based prototype together with a number of industrial partners.