Search the digital library catalog

Abstract: In response to the growing demand for accurate freight forecasting in sea-rail intermodal transportation, particularly under the constraints of stringent data protection regulations, we introduce a privacy-preserving, AI-based framework that focuses on the micro-level identification of container transport potential. The framework combines Vertical Federated Learning (VFL) with advanced feature and sample selection techniques. It leverages privacy-preserving methods, such as homomorphic encryption and random noise, enabling secure collaboration between ports and railways while safeguarding commercially sensitive data. Through extensive experiments, our framework demonstrates superior performance in predicting container transport demand, significantly improving the accuracy of resource allocation and scheduling decisions for rail operators. The framework not only ensures compliance with data protection regulations but also provides valuable insights into intermodal transportation planning, optimizing both railway operations and customer service quality. This approach offers a practical solution for improving strategic decision-making in the sea-rail intermodal sector amid increasing privacy demands and complex logistical challenges.
Keywords: freight demand forecasting, container transportation demand forecasting, vertical federated learning, privacy-preserving methods, sample and feature selection, machine learning, homomorphic encryption, resource allocation and scheduling
Published in DKUM: 20.01.2026; Views: 0; Downloads: 0
Full text (1,37 MB)
This document has many files! More...
This document is also a collection of 1 document!

Abstract: This study employed a recursive mixed-process model to analyze how sociodemographic characteristics affected household transportation expenditures and car ownership in Taiwan. Transportation expenditures were segmented into those for private vehicle use and those for public transport services. Data on households were sourced from Taiwan’s Family Income and Expenditure Survey for the years 2002 and 2022. The principal findings were as follows. First, household travel demand varied by household life cycle stage, with middle-aged households exhibiting the highest travel demand. Older households also exhibited substantial travel demand and had higher transportation expenditures and car ownership rates than households headed by individuals under 29 years old did. A finding of increased mobility among households headed by older adults reflected longer life expectancy, improved health, and greater wealth. Second, household composition considerably affected transportation expenditures and car ownership. For example, additional family members were typically associated with increased transportation expenditures. However, additional members aged 14 years or younger were associated with reduced public transportation expenditures because private vehicles often replaced public transit because they were used for caregiving and delivery. By contrast, additional older family members (aged 65 years or older) were associated with reduced private transportation expenditures and increased public transportation expenditures, reflecting older individuals’ limited wealth and the high costs of private vehicle use. These results clarify the determinants of transportation expenditures and highlight the characteristics of Taiwanese households reliant on private vehicles. As family structures change and population aging continues, age-friendly public transportation systems should be prioritized in the development of transportation.
Keywords: car ownership, transportation expenditures, population aging, travel demand
Published in DKUM: 28.10.2025; Views: 0; Downloads: 2
Full text (765,14 KB)
This document has many files! More...

Abstract: The study of commuting behavior at the University of Maribor (UM) was the subject of our research, which focused on the building complex of the four technical faculties (BCTF) and was based on the analysis of two questionnaire surveys (with 1057 and 462 respondents, respectively) and the transport policies implemented at the study site from 2010 to 2020. The research aimed to identify the factors influencing student and staff mode choice/shift over a decade period and to understand the weaknesses, strengths, and opportunities for improving sustainable mobility at the university. Since 2010, active commuting has predominated among students, while car use has decreased by 22%. Female students were 16% more likely to walk than their peers, while male students were 5% and 12% more likely to use bicycles and cars, respectively. Active commuting and car use by staff have not changed since 2010, and there was an insignificant difference between genders, 63% of whom used cars. Mode shifts were primarily related to trip origins, subsidization of bus use, availability or unavailability of free parking, and parking fees. Questionnaire responses were a powerful tool for finding the most effective interventions to manage transport at universities. The results also suggest that transport policies can be more effective if they are planned in coordination with housing policies.
Keywords: sustainable commuting, travel behavior, modal choice, active transport, effective interventions, transport demand management, university
Published in DKUM: 14.03.2025; Views: 0; Downloads: 6
Full text (1,71 MB)
This document has many files! More...

Abstract: This paper presents a real-time wearable system designed to assist Parkinson’s disease patients experiencing freezing of gait episodes. The system utilizes advanced machine learning models, including convolutional and recurrent neural networks, enhanced with past sample data preprocessing to achieve high accuracy, efficiency, and robustness. By continuously monitoring gait patterns, the system provides timely interventions, improving mobility and reducing the impact of freezing episodes. This paper explores the implementation of a CNN+RNN+PS machine learning model on a microcontroller-based device. The device operates at a real-time processing rate of 40 Hz and is deployed in practical settings to provide ‘on demand’ vibratory stimulation to patients. This paper examines the system’s ability to operate with minimal latency, achieving an average detection delay of just 261 milliseconds and a freezing of gait detection accuracy of 95.1%. While patients received on-demand stimulation, the system’s effectiveness was assessed by decreasing the average duration of freezing of gait episodes by 45%. These preliminarily results underscore the potential of personalized, real-time feedback systems in enhancing the quality of life and rehabilitation outcomes for patients with movement disorders.
Keywords: Parkinson’s disease, freezing of gait, machine learning, real-time systems, wearable devices, on-demand stimulation
Published in DKUM: 31.01.2025; Views: 0; Downloads: 12
Full text (6,29 MB)
This document has many files! More...

Keywords: flexibility, baseline, demand response, distribution transformer, congestion management, power flow control, peak shaving, load shifting, predictive models, machine learning
Published in DKUM: 05.01.2024; Views: 386; Downloads: 66
Full text (9,03 MB)
This document has many files! More...