The surge in demand for urban logistics has positioned unmanned aerial vehicles (UAVs) as a vital solution for improving last mile delivery efficiency in complex urban environments. To overcome the ground safety bottleneck in the large-scale application of logistics UAVs, this paper establishes a full-chain framework of “risk characteristics–quantitative model–risk cost” that proposes an innovative dual-state risk–cost optimization model based on aircraft-cargo separation probability. The study quantifies the separation probability using the Weibull distribution, corrects the risk value assessment deviation under the traditional single fall assumption, and establishes a risk probability model for the separation state by combining the dynamic characteristic differences between the UAV body and cargo, enabling the generation of rasterized risk maps by integrating spatial parameters, including population density and shelter coefficients, to implement risk–cost conversion and support airspace management. Case verification reveals that in areas with different land uses, there are significant differences in the ground risk values and risk costs of logistics UAVs. Additionally, introducing the aircraft-cargo separation rate improves risk assessment accuracy by 57%, and the ground risk in the aircraft-cargo integrated state is 4.31 times higher than in the separated state. This study provides core theoretical support for the safe operation and route decision-making involving logistics UAVs in complex urban scenarios.

Private vehicle overreliance in Klang Valley, Malaysia contributes to severe traffic congestion, air pollution, and carbon emissions, yet public transport modal shift remains far below national policy targets. This study extends the Theory of Planned Behaviour (TPB) by incorporating two external constructs, environmental concern (EC) and technology adoption (TA), to investigate the behavioural intention of 483 private vehicle users to switch to rail transit. Hypotheses were tested using partial least squares structural equation modelling (PLS-SEM). The results show that, with the exception of subjective norm ($\beta =0.064$, $p=0.180$), all constructs exert significant positive effects on behavioural intention. Technology adoption emerged as the strongest direct predictor ($\beta =0.417$), followed by perceived behaviour control ($\beta =0.340$). Environmental concern operated exclusively through indirect pathways, mediated by attitude and perceived behaviour control, and exhibited the largest effect size on the TPB components ($f^2$ = 0.380–0.449, large effect).

Existing timetable optimisation models for urban rail transit predominantly adopt operator-oriented objectives with assumed passenger-related weights. This paper proposes a passenger satisfaction-oriented timetable optimisation framework in which satisfaction weights are empirically derived from confirmatory factor analysis and Cramér’s V analysis of survey data collected from 439 passengers on Nanning Rail Transit Line 1. Seven scheduling-related service attributes are formally expressed as functions of timetable decision variables, establishing a direct linkage between passenger perception and scheduling decisions. A multi-objective model minimises a weighted combination of passenger dissatisfaction, operational cost, and stranded passenger ratio, solved by a Passenger Satisfaction-oriented Adaptive Dispatch Heuristic (PS-ADH) integrating simulated annealing with a passenger flow simulation module. Case study results demonstrate simultaneous improvements of 3.78% in composite objective value, 3.25% in passenger dissatisfaction, and 3.25% in operational cost, with a 27.4% reduction in stranded passengers. The optimised strategy is selected consistently across all ten random initialisations (CV = 0.13%). Sensitivity analysis reveals a structural break at cost weight $\beta =0.4$, beyond which the optimal strategy shifts qualitatively toward cost minimisation at the expense of service quality. The framework provides a transferable methodology for integrating passenger perception data into rail transit scheduling for emerging urban rail systems.

Passenger satisfaction in urban rail transit is shaped by operational service attributes whose relative importance may shift across operating conditions. This study investigates the context-dependent associations between operational service attributes and passenger satisfaction under normal and disrupted service conditions, using Nanning Rail Transit Line 1 as an empirical case. Based on 439 valid survey responses, ordinal logistic regression, confirmatory factor analysis, and non parametric statistical tests were employed to identify key satisfaction drivers across two operating scenarios. The results reveal that train frequency is the strongest independent predictor of overall satisfaction under normal operating conditions, while the number of stops displaces train frequency as the dominant predictor under service disruptions. Alighting time, non significant under normal operations, becomes independently significant under disruption, alongside platform waiting time which remains significant across both conditions, reflecting the heightened salience of congestion management when service stability is compromised. Exploratory segment analysis further suggests that satisfaction may vary by passenger characteristics, with nominally significant differences observed for leisure users with respect to coach congestion, work commuters with respect to running speed under disruption, and high-frequency riders with respect to train frequency, though these patterns did not survive correction for multiple comparisons. These findings demonstrate that the relative importance of operational service attributes is condition-sensitive and potentially user-specific, implying that a single uniform operational strategy is insufficient across varying service states. Transit agencies should prioritise service frequency and headway consistency under normal conditions, while shifting emphasis to maintaining service coverage and managing passenger flow during disruptions.