The study proposes an interpretable machine-learning approach for predicting bus arrival delays based on the Automatic Vehicle Location (AVL) system. The data describe vehicle movements and schedule adherence at successive stops, enabling delay estimation both along the route and across different time periods. Delays are defined as deviations between observed arrival times and scheduled times at individual stops, allowing both spatial and temporal delay patterns to be examined. The approach integrates ensemble machine-learning models with nested, temporally aware cross-validation and hyperparameter optimization. This validation strategy preserves the temporal order of observations, ensuring that model evaluation reflects realistic forecasting conditions and avoids information leakage. In particular, training and test sets are separated along the time axis, preventing future observations from influencing model training. Three ensemble models—Random Forest, XGBoost, and CatBoost—achieved comparable accuracy (RMSE ≈ 5.4 min; MAE ≈ 3.6 min; R² ≈ 0.44). Model performance was evaluated at the level of individual stop arrivals, reflecting short-term delay prediction under operational conditions. Comparable performance across models indicates that the observed patterns are not specific to a single modelling technique. This consistency suggests that the identified delay patterns are robust with respect to model choice rather than driven by algorithm-specific assumptions. The results show that the proposed approach enables the identification of combined temporal and spatial feature interactions associated with bus delay magnitudes within the analysed case study. These interactions mainly involve time-of-day effects and stop sequence, illustrating how delays accumulate as vehicles progress along the route. This pattern reflects the cumulative nature of delays in mixed-traffic operations, where early disturbances propagate toward downstream stops. SHAP values were used to provide a quantitative interpretation of these interactions and to identify the most influential predictors in the analysed models. The analytical structure, designed around weekly data partitions, supports delay monitoring and operational analysis. The results indicate that the proposed approach can support the analysis and interpretation of bus delay patterns within similar operational contexts.
Dhaka’s first mass rapid transit (MRT) system has marked a critical shift in the city’s public transport landscape, aiming to provide high-capacity, and reliable mobility. However, the effectiveness of this transportation mode depends on seamless first and last mile (FLM) connectivity. This study examines whether paratransit users differ meaningfully from active mode travelers in FLM connectivity for MRT system where paratransit like rickshaw is predominant. Using 1,513 valid home-interview responses, the research applies Principal Component Analysis (PCA) to extract three latent constructs, i.e., Travel Utility Perception, Travel Time & Cost Impact, and Socio-Demographic Profile. Structural Equation Modeling (SEM) is used to assess distinct models for walking and rickshaw as FLM modes. The findings reveal that the behavioral factors influencing paratransit use are closely aligned with those of active transport users, implying that in compact urban settings, paratransit may function as a substitute for active travel, potentially due to the lack of adequate walking and cycling infrastructure. Two models are then combined, referring to active FLM modes to investigate how their trip making behavior interact with each other. Key findings indicate that trip distance, fare sensitivity, time savings, and socio-economic status significantly influence FLM mode choice. The combined model of active travel yielded the best fit (RMSEA = 0.102; CFI = 0.848; TLI = 0.875), suggesting rickshaws function as pragmatic extensions of active travel rather than discrete modes. These results emphasize a continuum-based strategy for FLM planning that integrates rickshaw-friendly and walkable infrastructure to maximize accessibility and inclusivity along MRT corridors in built-up areas.
Disbanding of freight trains on sorting humps is a key process in railway transport. At most stations, this process still involves human operators controlling retarders and car speed. The lack of models that account for human participation limits both the reliable evaluation of hump automation measures and the application of computer-based train release planning tools. The purpose of this study is to improve the methods of modelling the disbanding trains on sorting humps to take into account human influence. The method of simulation modelling was used as the main research method. Determining the speed and time of rolling of the hitches is carried out by solving the differential equation of motion where distance is used as a variable. Setting the indicators of the process of disbanding of trains is carried out on the basis of the statistical processing of a series of calculation experiments on the rolling of cuts. In order to take into account the human participation in the process of disbanding the trains in the simulation model, additional restrictions are set on the choice of braking modes of the cuts, which are associated with the need to switch the attention of the operator during the simultaneous control of several braking positions, as well as with the transition of the controller between different tracks. The scientific novelty of the work consists in the improvement of the model of the disbanding of trains on the sorting humps, which, unlike the existing ones, allows to take into account the influence of operators of brake positions and speed regulators of cars on the indicators of the sorting process. The practical significance of the work lies in improving the evaluation of automated train release control systems and supporting the development of decision-support tools for train release planning under uncertainty in cut rolling characteristics and braking execution.
The following paper presents research aimed at identifying the most critical risks and their mitigations in Urban Air Mobility (UAM) operations. This topic is one of aviation's most significant challenges in the coming decades. Having many flying vehicles in a single airspace requires an innovative approach, rule redefinition, and traffic management. Some solutions are scalable and can be adapted from general aviation. Therefore, stakeholders must address new risks and implement dedicated methods while maintaining the highest level of operational safety. Simulation research is needed to validate solutions before systems operate in real environments. The response to those challenges is the development of a simulation tool that can serve as a test benchmark. The study is divided into two sections: identifying potential risks associated with the rapidly growing UAV market and its applications in urban environments and developing a simulation tool that addresses various Urban Air Mobility challenges. A set of test cases is presented to demonstrate the tool’s functionality and capabilities for further analysis. The paper reviews the United States and European Union approaches to UAM integration, including NASA, FAA, SESAR, and EASA initiatives, and highlights differences in operational concepts and regulatory frameworks. The research identifies major categories of risks related to UAV operations, including technical failures, environmental hazards, human factors, and cybersecurity threats. Long-term challenges associated with increasing traffic density, autonomous operations, and airspace organization are also discussed. The research evaluates scalable safety solutions derived from commercial aviation and analyzes urban airspace concepts such as layers, zones, sky-lanes, and sky-corridors. The developed simulation environment, implemented for the Warsaw metropolitan area, enables modeling of large-scale UAV and VTOL operations, no-fly zones, vertiport hubs, and traffic distribution. The results demonstrate the importance of dedicated traffic structures, altitude separation, and decentralized traffic management systems in ensuring safe and efficient Urban Air Mobility operations.
The rapid electrification of road transport is reshaping the automotive market, particularly in Europe, where battery-electric vehicles (BEVs) are becoming increasingly prevalent. Evaluating vehicle-level technological performance is critical for understanding competitiveness in this evolving landscape. This study aims to systematically assess and rank 174 BEV models in the B and C segments, considering technical and performance criteria, to identify technologically competitive vehicles and to compare the technological performance of Chinese manufacturers with that of established global brands. Eight technical indicators were selected to capture key dimensions of competitiveness: driving range, battery capacity, energy efficiency, fast-charging speed, maximum speed, trunk capacity, acceleration, and full-charging time. An integrated multicriteria decision-making (MCDM) framework was applied, combining Shannon’s entropy, the CRITIC method for objective weighting, the TOPSIS method for ranking, and the Borda count for consensus aggregation. The proposed MCDM framework offers a replicable and robust approach for benchmarking technological performance across heterogeneous BEV models. The analysis reveals that European manufacturers, particularly German brands, continue to dominate the highest-ranking positions due to well-established engineering capabilities, long-term investment in innovation, and patent-based technological leadership. Chinese producers included in the sample, while representing a growing share of the market, display strong performance in battery technology, cost efficiency, and manufacturing scale. Yet, their presence in the European B and C segments remains limited. The results further highlight that BEV competitiveness is multidimensional, reflecting trade-offs between performance, efficiency, and usability. C-segment vehicles tend to prioritise extended driving range and larger battery capacity, whereas B-segment models focus on energy efficiency and compact design, catering to urban mobility requirements. The study demonstrates that a single attribute cannot determine technological competitiveness; rather, it depends on a balanced combination of multiple criteria. These findings provide policymakers, manufacturers, and investors with insights into the technological strengths and weaknesses of contemporary BEV models, supporting evidence-based discussions on sustainable mobility and industrial development.
