Train-ride quality evaluation of the Elizabeth Line using machine learning

This present study raises the possibility that our new AI model can embrace crowdsensing and decentralize live monitoring of train ride quality in three-dimensional space (considering both linear and rolling accelerations), as clearly demonstrated. This study has also demonstrated the robustness of the ML model for applications to both surface and underground trains. By data assimilation and automated retraining, the AI model can fully address and determine the ride quality taking into account both ride comfort and rolling motions.