A spatiotemporal fusion approach to track quality index construction and irregularity prediction

Feng Wang, Hao Cheng, Y Cao, Guanyu Hu, F Wang, Xiao Ma, Jiao Li

Abstract Track irregularity reflects deviations in track geometry and serves as a key indicator of rail-way infrastructure performance. Accurate evaluation and prediction of track quality are cru-cial for detecting potential safety risks and guiding maintenance decisions. Traditional Track Quality Index (TQI) models use fixed weights and often fail to capture dynamic track evolu-tion. To address this, we propose an Optimized Track Quality Index (OTQI) framework for adaptive assessment and prediction. By incorporating temporal degradation monotonicity and spatial heterogeneity among track segments, a spatiotemporally constrained optimiza-tion model is formulated to adaptively adjust TQI component weights, enabling OTQI to more accurately represent actual track health and evolution trends. A Transformer-based prediction model is further introduced to forecast the degradation trajectory of OTQI. Ex-perimental results show that the optimized OTQI strongly correlates with empirical assess-ments and, when combined with the Transformer model, substantially improves prediction accuracy and robustness. The proposed framework presents a refined approach for track quality assessment and prediction, and provides robust technical support for data-driven pre-dictive maintenance of railway infrastructure.