Transportmetrica A-Transport Science最新文献

A speed perturbation model is proposed in this paper for the heterogeneous platoon to measure the interaction between vehicles and investigate the propagation laws of perturbations. Then, a ratio is defined to quantify the relationship between the communicating information impact (CII) and the car-following behaviour impact (CFI). A modified IDM is employed to evaluate the role of communicating information in perturbation propagation. Results show that communicating information can suppress the amplification of perturbation. Under the stable environment, the interaction between vehicles decreases with frequency. There is a critical frequency ${\omega }_c$ (value of 0.55 in this case) that distinguishes the relationship between the CII and CFI. When the frequency is smaller than ${\omega }_c$, the CFI is larger than CII; otherwise, the CII dominates the interactions between vehicles. Under the unstable environment, the interaction between vehicles increases first and decreases then, where the vehicles are mainly affected by the car-following behaviours.

Bike-sharing improves individual mobility, considerably reshaping the landscape of job accessibility and commuting time. Existing empirical studies in urban transportation involving commuting usually collect survey data at the aggregate level. A comprehensive understanding of the influence of bike-sharing on commuting and job accessibility at the city level is still missing in developing countries. Using mobile phone data in Beijing, this study addresses these questions with a commuting mode model and cumulative accessibility model. The results indicate that bike-sharing could lead to a decrease in commuting time and an increase in job accessibility. The availability of bike-sharing services has a positive relationship with its effectiveness. Meanwhile, bike-sharing significantly reduces the horizontal and vertical inequality in commuting time and job accessibility at both the individual and spatial levels. These findings provide insights into the popularity of bike-sharing in China, shed light on the equity influence of bike-sharing, and provide a quantitative measurement of the benefit of bike-sharing.

Emergency Medical Services (EMS) constitute a crucial pillar of today's cities by providing urgent medical responses to their citizens. Their study is often conducted via simulation, as the assessment of planning decisions is generally unfeasible in the existing systems. However, such models can become computationally expensive to run. Thus, metamodels can be used to approximate the simulation results.

In this work, a simulation metamodelling strategy supported on an active learning scheme is proposed to analyse the survival rate of a simulated EMS. The exploration process is guided through a series of grids towards simulation input regions whose output results match a specific survival rate defined a priori. This provides an efficient way of exploring the search space by channelling the computational effort to the most important input values, supporting the advantages of these methodologies in the EMS field, where their application is still seldom to the best of our knowledge.

The second-best tradable credit scheme is studied to analyse its effects on travel mobility by a bilevel programming model. The transport authority optimises the credit charging scheme to minimise the total system travel time and travellers choose paths to minimise their generalised travel costs. A genetic algorithm with some improvements on the selection operator is used to solve the proposed bilevel programming model. In numerical experiments, we discuss the implications of travellers’ path choice behaviour and traffic flow distribution under the second-best tradable credit scheme with different combinations of credit-charging links. The results show that the credit charging scheme with all links is obviously better than other schemes in the total system travel time. Besides, the credit charging scheme using congested links or duplicate links of the maximum traffic flow path as credit-charging links is effective in reducing traffic flow on congested links and mitigating the congestion of the whole transportation network.

The paper aims to develop an improved Fog-related Intelligent Driver Model (FIDM) that reproduces drivers’ car-following behaviour features by taking into account unobserved driver heterogeneity in fog condition. A multi-user driving simulator experiment was performed, and a vehicle fleet consisting of nine vehicles was tested in different fog and speed limits conditions. The experimental results showed that the unobserved driver heterogeneity (the combination of intra-driver heterogeneity and inter-driver heterogeneity) tended to increase as the fog density decreased. The average following distance tended to increase with the decrease of fog density and increase of speed limit. Two indexes were proposed to verify the performance of the FIDM. The results showed that FIDM performed better in reproducing unobserved driver heterogeneity and average following distance compared to the current popular car-following models. This study contributes to an improved car-following model for better understanding traffic flow phenomena under foggy conditions.

The data required for transportation applications can be retrieved from mobile phones without the necessity of additional infrastructure. Thus, we propose a procedure that involves two stages – data preprocessing and transportation mode detection – for detecting the transportation mode (i.e., car and bus) on the basis of sighting data. In the data preprocessing stage, two detection rules are used for eliminating oscillations that occur when a mobile phone intermittently switches between cell towers instead of connecting to the nearest cell tower. In the transportation mode detection stage, two supervised machine learning methods, namely support vector machine (SVM) and a deep neural network (DNN), are used to detect transportation modes. Experimental results indicated SVM achieved a higher accuracy (96.49%) in transport mode detection than did the DNN (69.65%) during peak hours. Moreover, travel time and starting time of a trip were identified as critical features affecting the accuracy of transportation mode detection.

To adapt to human-driving habits, this study develops a personalised car-following model via a memory-based deep reinforcement learning approach. Specifically, Twin Delayed Deep Deterministic Policy Gradients (TD3) is integrated with a long short-term memory (LSTM) (abbreviated as LSTM-TD3). Using the NGSIM dataset, unsupervised learning-based clustering and data feature analyses are performed. The driving characteristics related to safety, efficiency and comfort are extracted for different driving styles, i.e. aggressive, common and conservative. Then, reward functions are constructed for different driving styles by incorporating their driving characteristics. By resorting to the TD3 policy within a recurrent actor–critic framework, LSTM-TD3 optimises the car-following behaviour via trial-and-error interactions according to the reward functions. Results show that compared with LSTM-DDPG and DDPG, LSTM-TD3 reproduces personalised car-following behaviour with desirable convergence speed and reward. It reveals that LSTM-TD3 can reflect the essential difference in safety, efficiency and comfort requirements among different driving styles.

A concept called Combined Alternate-Direction Lane Assignment and Reservation-based Intersection Control (CADLARIC) was recently proposed for better management of directionally unrestricted traffic flows in a Connected and Autonomous Vehicles (CAVs) environment. In CADLARIC, vehicles must position themselves in a proper lane before they reach the downstream intersection, which enables the resolution of vehicular conflicts both between intersections and, as traditionally, within the intersection boxes. CADLARIC has shown very promising results, but it is quite infrastructurally demanding, requiring six lanes per intersection approach. To overcome this problem, we propose here Combined Flexible Lane Assignment and Reservation-based Intersection Control (CFLARIC), a more robust concept that offers a full spectrum of lane assignment possibilities in combination with the appropriate reservation-based intersection control. Three distinctive CFLARIC strategies are tested on a simulated three-intersection corridor from West Valley City, Utah. The efficiency and safety performance of the proposed CFLARIC scenarios are evaluated through a comparison with Fixed-Time Control (FTC) and Full Reservation-based Intersection Control (FRIC), both with conventional lane assignments. The results illustrate that CFLARIC scenarios: (i) outperform FTC and FRIC in terms of efficiency (delay and number of stops), and (ii) improve overall safety (by reducing a number of conflicting situations) when compared to FRIC. The findings of this study prove that flexible control concepts such as CFLARIC have a great potential to improve safety and efficiency in future CAV environment. Future research needs to define under which conditions and how the proposed concept could be partially implemented in field operations.

Growing research attention is focusing on Automated Vehicle (AV) technologies, promising significant safety benefits. An in-depth understanding of human driving will play an important role in determining the most acceptable AV behaviour, supporting passenger comfort and thus the adoption of the technology, but also the optimal prediction of the behaviour of the surrounding traffic. The current study examined 1740 real-world motorway pull-out manoeuvres (pull-out distance, speed differential with the leading vehicle, manoeuvre duration, pull-out comfort zone) under different conditions. The results highlighted the significant impact of the surrounding traffic and the driving characteristics on or before the manoeuvre initiation point, which reflected the overtaking strategy selected. The findings can inform the design of automated overtaking systems that resemble human driving and thus encourage their uptake; in addition, they can assist the intention prediction for lane keeping assistance systems in order to optimise the system’s response to cutting in and pull-out manoeuvres.

Transfer synchronisation aims at reducing the waiting time of transferring passengers to achieve a seamless transfer. Taking control at the operation level is widely regarded as an effective method to make real-time response to transfer synchronisation. However, the irregularity of bus line service after transfer may be caused due to the real-time control before the transfer behaviour. In this paper, we propose a two-stage speed control strategy to improve the transfer efficiency and the regularity of bus service after transfer simultaneously. The influence of speed control is estimated by constructing dynamic rolling horizon. Considering the service performances of both the transfer node and the next stop, we develop a multi-objective optimisation model. Experimental results show that compared with the one-stage control, the two-stage control can significantly reduce the deviation of headway after transfer while reducing the transfer time.