Machine learning (ML) is a crucial component of artificial intelligence that has recently attracted attention for its application in logistics. ML algorithms are used on large datasets. They create logic correlations among given data and provide predictions of specific values. This research paper aims to conduct a systematic literature review to showcase the potential applications of machine learning in urban logistics systems, specifically focusing on enhancing satisfaction for postal logistics operators and their customers. The authors used various research publication databases in this context (Web of Science, Scopus, Google Scholar etc). The analysis of different models provides insights into diverse aspects, such as predicting product prices and types of cargo, evaluating user satisfaction, forecasting user departures, assessing optimal geographical locations for implementing postal centres, predicting purchase times before online orders, estimating delivery times in the last phase of the logistics chain and more. The significance of this research is highlighted through the identification of shortcomings in existing literature, offering guidelines for future research in developing new machine learning model for optimal operator selection. This model aims to achieve improvements in both customer and operator satisfaction simultaneously.

With the potential for fast, contactless and environmentally friendly delivery, unmanned aerial vehicles (UAVs) have gained increasing attention and application due to their cost-effectiveness and convenient and rapid delivery operations. In future cities, a multi-level airport that supports vertical take-off and landing (VTOL) of UAVs and forming a delivery network is necessary to improve delivery efficiency and provide a competitive advantage. This paper proposes a multi-level airport location-routing problem for UAVs that considers UAV flight energy consumption and operational costs. The goal is to minimise the number of locations and minimise delivery path planning while meeting delivery demands within the service range. Based on the traditional distribution centre site-path problem, the UAV distribution network is constructed to solve the problem of airport location and flight path planning, and the two-layer genetic algorithm is used to solve it. Based on this, the validity of the model and algorithm is verified using the urban area of Tianjin as an example. The experimental results show that the constructed model can be used for UAV airport layout planning, which is applicable to large-scale, multi-aircraft-type and multi-level airport layout planning. Data analysis results indicate that when the location layout of the vertical hub airport is on the edge of the VTOL points, both the flight distance and the total cost of the delivery network relatively increase. Increasing the payload capacity will reduce the number of UAV operations, but the total cost shows a decreasing-then-increasing trend. This study can provide a theoretical basis for the selection of airport sites and UAV types in future UAV urban delivery networks.

With the ongoing urbanisation, the subway has become a vital component of modern cities, catering to the escalating demands of a mobile population. However, the increasing complexity of passenger flows within subway stations poses challenges to operations management. To optimise subway operations and enhance safety, researchers focus on extracting and analysing pedestrian trajectories within subway stations. Traditional trajectory extraction methods face limitations due to manual feature design and multi-stage processing. Leveraging advancements in deep learning, this paper integrates M-DeepSORT with YOLOv5 and proposes a feature association matching approach that addresses trajectory drift issues through simultaneous consideration of motion and appearance matching. The confidence-based (CB) Kalman filtering method is proposed to address the issue of random noise in pedestrian detection within subway scenes. The introduction of a momentum-based passenger trajectory centre update method reduces jitter, resulting in smoother trajectory extraction. Experimental results affirm the effectiveness of the proposed algorithm in detecting, tracking and statistically analysing subway station corridor passenger flow trajectories, demonstrating robust performance in diverse subway station scenarios.

The highway tunnel plays a critical role in highway traffic flow, yet its sections are particularly susceptible to traffic accidents. The research shows that the safety measures in the tunnel have a certain effect on improving the safety in the tunnel, but there is a lack of evaluation methods for the use effect of safety measures in the tunnel. To study the application effect of safety measures in tunnels (mainly strobe lights and information boards), this paper takes the driver’s subjective feelings and vehicle speed changes as indicators to evaluate the application effect of safety facilities. The Xingshuliang Tunnel in Shaanxi Province, which has been operated and meets the test standards, is used as the test site, and the driver between the Yaozhou and Huangling sections is randomly selected as the test object for data collection. Subjective feelings are mainly obtained by social survey methods to obtain data samples, and the driving speed is collected by NC2000, non-contact five-wheel instrument, video recorder and other equipment. The statistical analysis method is used to study the driving speed of each section inside and outside the tunnel and the driver’s response. According to the changing trend of speed, the weight of each test section is calculated by the combination of analytic hierarchy process and quantitative statistical method, and the comprehensive influence degree of safety measures is evaluated. The results show that both the strobe light and the information board induce the driver to reduce the driving speed by 3.1%, which can effectively reduce the driving speed. The strobe light mainly acts on the tunnel entrance and the inside of the tunnel, with a maximum influence range of 205 m. The information board has the greatest effect at the tunnel entrance, with a maximum influence range of 200 m. The above results provide a useful reference for the arrangement of safety measures and put forward the arrangement method of tunnel safety measures in combination with the conclusion, to help improve the safety of the driving environment in the tunnel.

Walking prevents disease and keeps older adults healthy. Studying the walking decision factors of ageing people is an imperative step in understanding and promoting their behaviour. The theory of planned behaviour (TPB) and the prototype willingness model (PWM) are two well-known frameworks that deal with reasoned and social reaction processes in decision-making. This paper used these frameworks to examine the motivational mechanisms of walking among older Chinese adults living in small towns and used the structural equation model (SEM) for regression analysis. The analysis was based on 407 questionnaires and compared two models. Model 1 is TPB and model 2 combines TPB with PWM. The results show that behaviour willingness (BW) is significantly correlated with behaviour intention (BI), and model 2 explains a higher proportion of intention variance than model 1. Perceived behavioural control (PBC) is the most significant predictor in the two models, which implies that walking usefulness and walking feasibility are critical to older adults’ willingness to walk and walking program development. Finally, the utility of the integrative model is discussed, in terms of the theoretical contribution to walking among older adults and the applied implications for the promotion of walking.

This study focuses on understanding the effects of shared mobility on travel behaviours and transport energy in the university campus. Using survey data collected from college students in Ningbo, China, a substitution model was developed to identify changes in travel modes with the introduction of shared mobility on college campuses and to quantify its impact on net energy saving. Considering the average time travelled and the life cycle energy unit of the trip, the before-and-after analysis was conducted to determine the travel behaviours and related transport energy of college students in 2016 and 2019. Compared with the data of 2016 when no shared mobility was introduced, 2019 data revealed three changes in travel behaviours. First, although the total number of trips per person decreased slightly, the trip distance increased in 2019. Second, the energy for trips by each student increased by 25% from 19,809 KJ in 2016 to 24,897 KJ in 2019. Third, the overall energy efficiency of the trips decreased. In conclusion, the effect of shared mobility introduced in the university campus on reducing the transport energy of college students has not been satisfactory.

The rapid expansion of the metro network, driven by urbanisation and a heightened focus on environmental sustainability, underscores the need for efficient and sustainable public transportation systems. This study utilises the West Midlands Metro system as a case study to investigate operational efficiency and utilisation challenges that are common across metro networks globally. Employing advanced simulation modelling with SIMUL8, this research evaluates the existing timetables and utilisation rates of the West Midlands Metro to uncover inefficiencies and untapped potential. Various scenarios, including increased service frequencies and disruptions at high-traffic stations, were simulated to provide actionable insights for optimising metro operations. Findings revealed that increasing service frequency from every 10 minutes to every 5 minutes enhanced utilisation levels and boosted the total number of completed services. Meanwhile, disruptions at major stops resulted in a reduction in utilisation in a negligible range. These results demonstrate that improved service frequency significantly bolsters operational efficiency and showcases resilience to disruptions with minimal impact on overall performance. As to future research, the study suggests that implementing adaptive scheduling through AI-driven maintenance and infrastructure improvements can further elevate the efficiency and passenger experience of metro operations.

This study asserts that paired aircraft can withstand specific wake turbulence levels and explores the longitudinal collision risk in closely spaced parallel runway approaches. The goal is to enhance the safety margin of the paired approach and allow for more flexible implementation. Based on QAR data, a theoretical spacing model for paired aircraft and a probability distribution of acceleration error are established to facilitate the analysis of the actual spacing of paired aircraft. Wake turbulence attenuation is modelled using large eddy simulation, creating a vortex attenuation model. Drawing inspiration from the Hallock-Burnham vortex model, new models for induced velocity and vortex core motion are proposed. The study assumes that trailing aircraft can handle certain wake intensities, leading to a new model for calculating wake turbulence safety intervals, limiting the trailing aircraft’s maximum roll angle to its critical limit. Using probability theory, a model for longitudinal collision risk is formulated, combining wake turbulence safety separation and the actual separation of paired aircraft. The study also examines various factors influencing longitudinal collision risk, emphasising the significant impact of crosswind conditions. It concludes that a stronger crosswind component reduces the wake turbulence safety separation, thereby increasing the risk of longitudinal collisions, particularly during the final stage of the approach. Notably, collision risk is directly proportional to the crosswind component and initial longitudinal separation, but inversely proportional to runway spacing.

This study introduces a novel, adaptable framework for identifying and prioritising road traffic accident hotspots using the Getis Ord Gi* spatial autocorrelation tool. The framework classifies regions as hotspots or coldspots based on accident severity and frequency. A unique weighting system is developed to compute the Crash Severity Index (CSI), considering the severity of crashes in terms of fatalities and injuries. The identified hotspots are prioritised using the CSI, providing policymakers with a structured approach to allocate resources for crash remedial measures. The main contribution of this work is the development of a flexible framework applicable to various cities, states or countries to improve road safety. The framework’s effectiveness is demonstrated through a case study in Punjab, India, revealing that Sangrur, Hoshiarpur and Police Commissionerate Ludhiana are the top three hotspots. The study also offers a detailed analysis of crash statistics in Punjab, emphasising the severity of pedestrian crashes. This approach addresses the current lack of structured hotspot identification and prioritization strategies, marking a significant advancement in road safety management.

This paper presents a closed queuing network model to address bike queues in bike-sharing systems with finite docks. The model tackles issues of bike spillover and user attrition due to fully occupied docks and bike shortages at stations. The objective is to determine throughput rates and other performance metrics for these systems. To overcome computational challenges, we propose an approximation algorithm based on the developed model. Our analysis reveals intrinsic properties of bike-sharing systems with finite docks: (i) The effective system throughput rate increases with bike fleet size and eventually converges to a ceiling value. (ii) Adding more docks at stations can unnecessarily increase or even decrease the effective throughput rate. (iii) Under certain conditions, the system can reach a self-balancing state, avoiding bike surpluses or deficiencies at each station and maximising throughput. (iv) Users can successfully return bikes with a limited number of tries, provided there is at least one station on their route with a non-zero probability of having available docks. A small-scale artificial example and a case study demonstrate the accuracy and applicability of the approximation algorithm and the properties of the systems.

Previous studies have primarily focused on the effect of the built environment on ridership during weekdays and weekends. This paper aims to investigate the spatial heterogeneity of the effect of built environment factors on ridership at metro stations during National Day holidays. Beijing is divided into three zones from inner to outer areas. Taking metro station boarding and alighting ridership during National Dayholidays as the dependent variable, 13 built environment factors were selected as independent variables according to the “7D” dimension of the built environment. The recommended pedestrian catchments (PCA) combinations for the three zones in Beijing are 400 m_500 m_400 m by using the Multi-Scale Geographically Weighted Regression (MGWR) model. We investigated the effect of built environment factors on metro ridership and spatial heterogeneity. The influencing factors that have significant effects on both boarding and alighting ridership are building density, number of commercial facilities, bus lines density, number of entrance and exit, number of office facilities, mixed utilization of land and road density. The MGWR model results are helpful to propose targeted strategies for revitalising the built environment around metro stations.

With the acceleration of urbanisation and the rapid increase in road traffic volume, the scientific prediction of traffic accidents has become crucial for improving road safety and enhancing traffic efficiency. However, traffic accident prediction is a complex and multifaceted problem that requires the comprehensive consideration of multiple factors, including people, vehicles, roads and the environment. This paper provides a detailed analysis of traffic accident prediction based on multi-source data. By thoroughly considering data sources, data processing and prediction methods, this paper introduces the various aspects of traffic accident prediction from different perspectives. It helps readers understand the characteristics of different data and methods, the process of accident prediction and the key technologies involved. At the end of the paper, the main challenges and future directions in road crash prediction research are summarised. For example, the lack of efficient data sharing between different departments and fields poses significant challenges to the integration of multi-source data. In the future, combining deep learning models with time-sensitive data, such as social media and vehicle network data, could effectively improve the accuracy of real-time accident prediction.

Modern transportation planning, building and city management tendencies are based on smart and green sustainability. Sustainable mobility and traffic safety can be improved by raising awareness of gender diversity in travel behaviour. The city of Novi Sad (Serbia) has had a tradition of traffic surveys for about 50 years. All measures for improvement of traffic in the city are based on this research. Travel characteristics in the city regarding gender differences are the source of different data, which can be based on sustainable urban mobility plans, transport demand management and gender equality levels. This paper analyses mobility, modal split and distribution by travel reason as a function of gender. Results were gathered using travel diaries (collected by household surveys) in the Smart plan of Novi Sad. The discussion is based on a descriptive analysis of the basic travel characteristics as a function of gender and a comparison with a previous traffic study – Nostram. The findings of the research indicated that there is a greater prevalence of male employment and higher usage of passenger vehicles among men than among women. Women use alternative modes of transport (walking, public transport and bicycle). Women do not have precise peak hours, while trips are scattered even outside peak hours. Their trips are more related to daily duties than entertainment.

Establishing simulation models is a widely used and effective approach for analysing passenger flow distribution in urban rail transit systems. Recently, multi-agent and discrete event-based simulation models have shown exceptional performance in studying passenger flow information within urban rail transit systems. While simulations of passengers and trains often yield satisfactory results, few models capture the overall operational status of urban rail transit systems. The complex interactions among stations, trains and passengers make it challenging to integrate these elements into a unified system framework. In this paper, we introduce a triple simulation framework that integrates stations, trains and passengers as foundational elements to comprehensively simulate the entire urban rail transit system and observe overall passenger flow distribution. Experimental results demonstrate that our system surpasses existing advanced simulation models, achieving an accuracy rate of 88.44% with a tolerance for a 30% deviation. To further illustrate the effectiveness of our framework in analysing passenger flows, we conducted experiments using the Nanjing Metro AFC dataset, analysing passenger flow distributions at stations and on trains.

Crowdshipping has garnered increasing interest due to its potential benefits for various stakeholders. However, despite challenges in attracting crowdshippers, limited research explores their preferences, including socio-demographic factors and the practical challenges providers face when testing or implementing crowdshipping. This study aims to identify key factors influencing willingness-to-work (WTW) among potential crowdshippers, both in general and within business-to-business (B2B) and business-to-customer (B2C) contexts. Based on the literature review, this paper identifies 19 barriers influencing WTW and develops 22 corresponding enablers to address these barriers. Using a survey of 432 participants from Slovenia, the overall significance of these factors without differentiating business models was first assessed. Then, chi-squared automatic interaction detection analysis was applied to predict WTW in B2B and B2C contexts, identifying variations across these models. The disclosure of a mobile number emerged as the most influential predictor in both settings. Other notable differences in enablers and barriers were observed depending on the business model. These findings emphasise the need to consider business models in future preference analyses and provide a foundation for targeted recruitment strategies for crowdshippers.