Vessel trajectory prediction is important in maritime traffic safety and emergency management. Vessel trajectory prediction using vessel automatic identification system (AIS) data has attracted wide attention. Deep learning techniques have been widely applied to vessel trajectory prediction tasks due to their advantages in fine-grained feature learning and time series modelling. However, most deep learning-based methods use a unified approach for modelling AIS data, ignoring the diversity of AIS data and the impact of noise on prediction performance due to environmental factors. To address this issue, this study introduces a method consisting of temporal convolutional network (TCN), convolutional neural network (CNN) and convolutional long short-term memory (ConvLSTM) to predict vessel trajectories, called TCC. The model employs TCN to capture the complex correlation of the time series, utilises CNN to capture the fine-grained covariate features and then captures the dynamics and complexity of the trajectory sequences through ConvLSTM to predict vessel trajectories. Experiments are conducted on real public datasets, and the results show that the TCC model proposed in this paper outperforms the existing baseline algorithms with high accuracy and robustness in vessel trajectory prediction.

2024 (Vol 36), Issue 6

To reveal the speed control behaviour and manoeuvring characteristics of direct vehicles that stop-go through signalised intersections, a large-scale field driving test was carried out in Chongqing to collect vehicle data under natural driving conditions. The characteristics of speed, longitudinal acceleration rate and their two-dimensional correlation were analysed for deceleration and acceleration behaviour at signalised intersections. Further, a sensitivity analysis of the simulation model on measured data was done with the micro-traffic simulation experiment of a signalised intersection. The following were observed: (1) Drivers’ speed-selection behaviours become more concentrated with closer distance from the stop point. The transects ±25 m from the stop point are abrupt change points in the discrete nature of driver speed-selective behaviours. (2) Drivers’ desire to decelerate during the stop-go through signalised intersections is more robust, with the magnitude of pedal manoeuvres for deceleration behaviours being more intense than that for acceleration behaviours. (3) There is a nonlinear correlation between longitudinal acceleration rate and speed. The longitudinal acceleration rate increased with increase in speed and then decreased with the inflection point at 15 km/h. (4) The micro-traffic simulation’s acceleration rate model is sensitive to measured acceleration rate parameters. This study guides the parameter setting of speed, deceleration rate and acceleration rate models for microscopic traffic simulation and for parameter calibration of the car-following model.

2024 (Vol 36), Issue 6

As a critical component of urban transportation, metro systems demand rigorous passenger flow safety management. This study proposes a comprehensive decision-making analysis method for metro station passenger flow safety management by integrating the entropy weight and TOPSIS methods. It aims to develop an evaluation model that accurately assesses and ranks the safety management practices of metro stations. To achieve this, 17 indicators related to station scale, safety management equipment, safety or security measures, investment in safety management and the effects of passenger flow management are selected to form an evaluation indicator system. The entropy weight method is employed to allocate weights to these indicators, reflecting their interrelatedness and importance. Subsequently, the TOPSIS method is used to establish a decision model that calculates the closeness of each station’s management practice to an optimal plan, allowing for the ranking of different stations’ safety management practices. The algorithms are developed and optimised using MATLAB, enabling efficient calculation and analysis. A case study involving real metro stations is conducted to validate the feasibility and effectiveness of the proposed evaluation method. The results demonstrate that this model provides an accurate assessment of metro station passenger safety management and offers decision-makers clear directions for improvement.

2024 (Vol 36), Issue 6

The study comprehensively evaluates the safety of contraflow left-turn lane intersection, characterised by unique traffic operational features distinct from conventional intersections. The evaluation specifically focuses on the process of left-turning vehicles entering the receiving lane within the intersection. The vehicle arrival rate of left-turning vehicles is analysed to identify vertical conflict features in contraflow left-turn lane design. By subdividing lanes within the intersection, the study delves into the lateral displacement of left-turning vehicles to establish lateral conflict features. To quantify the overall conflict potential, a multiple unit conflicts index is derived by integrating both vertical and lateral conflict features. Furthermore, the double index left-turn conflict model is constructed by introducing the potential collisions severity index during the conflict process. The results indicate that conflict hotspots along the vehicle travel path are primarily concentrated in two regions: (1) at pedestrian crosswalks and within a 2-meter extension; (2) within a range of 6 to 18 meters from the pedestrian crosswalk. The proposed model demonstrates good evaluation effectiveness, providing valuable insights into enhancing the safety of contraflow left-turn lane intersections.

2024 (Vol 36), Issue 6

The aim of this paper is to highlight the vulnerability of Maritime Autonomous Surface Ships (MASS) to cyber-attack and to illustrate, through a simulation experiment on a testbed, how to mitigate a cyber-attack on the MASS thruster controllers during low-speed motion. The first part of the paper is based on a scoping review of relevant articles in the field, including some MASS projects, related cyber threats and modelling techniques to improve cyber resilience. In the second part of the paper, a cyber-attack on the MASS thruster controllers at low speed motion is illustrated along with the impact of the attack on the trajectory motion. The Kalman filter, as an additional device to the thruster controllers, is used as a cyber-attack mitigation aid. Under the conditions of a simulated intrusion on the input and output signals of the thruster, the experiments conducted in the MATLAB Simulink environment provide an insight into the behaviour of the MASS propulsion subsystem from the perspective of the low-speed trajectory, with and without the Kalman filter.

2024 (Vol 36), Issue 6

This paper focuses on the online energy-saving operation control problem for passenger and freight trains running in a single-track railway line. Firstly, we design a centralised optimisation method to generate energy-saving reference profiles for both passenger and freight trains, in order to improve the punctuality of passenger trains and to reduce the total running time of freight trains in a central way. Secondly, we propose the distributed model predictive control (DMPC) based online trajectory optimisation problems for both types of trains, subject to their respective operational constraints including safety, punctuality, static speed limits and temporary speed restrictions. Then we formulate an online train operation control algorithm based on the centralised optimisation method for the initialisation of train trajectories and the DMPC method for the online trajectory planning. Finally, the proposed algorithm is applied to case studies of passenger and freight trains in a single track railway, and the numerical simulation results show that the proposed algorithm can realise online control for energy-saving train operation in the presence of input disturbances and temporary speed restrictions.

2024 (Vol 36), Issue 6