Understanding why cities diverge in motor vehicle ownership trajectories is critical for designing differentiated and sustainable transport policies. This study develops an integrated national–city analytical framework to examine heterogeneous urban motorization processes in China. A national Gompertz curve is first estimated to represent the benchmark income–ownership relationship. City-specific deviations are then decomposed into two interpretable dimensions: a horizontal stage parameter (h), capturing relative advancement or delay in motorization timing, and a vertical scaling parameter (s), reflecting persistent ownership intensity differences conditional on income. Results show substantial and multi-dimensional heterogeneity across cities. Stage timing (h) and ownership intensity (s) are only weakly correlated, indicating that earlier transition into higher motorization stages does not necessarily imply above-benchmark ownership intensity. Random forest models with time-forward validation demonstrate strong explanatory power (R² ≈ 0.88 for h and 0.80 for s). SHAP-based interpretation reveals that stage deviation is primarily associated with transport supply and urban structural characteristics, whereas ownership intensity deviation is more strongly linked to urban spatial scale and economic structure. Regulatory measures, including purchase and driving restrictions, exhibit comparatively smaller and heterogeneous effects. By disentangling timing and intensity dimensions of urban motorization, this study refines conventional income-based diffusion models and provides quantitative evidence that structural urban characteristics play a more fundamental role than regulatory interventions in shaping inter-city motorization differences.

Accurate heavy-duty vehicle load estimation is crucial for transportation and environmental regulation, yet current methods lack precision in data accuracy and practicality for field implementation. We propose a Self-Supervised Reconstruction Heterogeneous Graph Convolutional Network (SSR-HGCN) for load estimation using On-Board Diagnostics (OBD) data. The method integrates physics-constrained heterogeneous graph construction based on vehicle speed, acceleration, and engine parameters, leveraging graph neural networks’ information propagation mechanisms and self-supervised learning’s adaptability to low-quality data. The method comprises three modules: (1) a physics-constrained heterogeneous graph structure that, guided by the symmetry (invariance) of physical laws, introduces a structural asymmetry by treating kinematic and dynamic features as distinct node types to enhance model interpretability; (2) a self-supervised reconstruction module that learns robust representations from noisy OBD streams without extensive labeling, improving adaptability to data quality variations; and (3) a multi-layer feature extraction architecture combining graph convolutional networks (GCNs) and graph attention networks (GATs) for hierarchical feature aggregation. On a test set of 800 heavy-duty vehicle trips, SSR-HGCN demonstrated superior performance over key baseline models. Compared with the classical time-series model LSTM, it achieved average improvements of 20.76% in RMSE and 41.23% in MAPE. It also outperformed the standard graph model GraphSAGE, reducing RMSE by 21.98% and MAPE by 7.15%, ultimately achieving < 15% error for over 90% of test samples. This method provides an effective technical solution for heavy-duty vehicle load monitoring, with immediate applications in fleet supervision, overloading detection, and regulatory enforcement for environmental compliance.