During daily work at a Transport Management Centre (TMC), the operators have to record and process a large volume of traffic information especially incident records. Their tasks involve manual classification of the data and then decide appropriate operations to clear the incidents on time. A real-time automatic decision support system can minimise an operator's responded time and hence reduce congestion. Besides standard descriptions (e.g. incident location, date, time, lanes affected), severity is an important criteria that operators have to evaluate based on all available information before any control commands can be issued. The NSW TMC and the research organisation Data61 in Sydney have collaborated to discover and visualise frequent patterns in historical incident response records, leading to the automatic classification of severity levels among past incidents using advanced machine learning, active learning and outlier detection techniques. The experiments were executed using 4 years TMC's incident logs from 2011 to 2014 which includes >40,000 records. The classification model achieved nearly 90% accuracy in five-fold cross-validation and is expected to help the TMC to improve its procedures, response plans, and resource allocations.

Advanced driving assistance systems (ADAS) have huge potential for improving road safety and travel times. However, their take-up in the market is very slow; and these systems should consider driver's preferences to increase adoption rates. The aim of this study is to develop a model providing drivers with the optimal trajectory considering the motorist's driving style in real time. Travel duration and safety are the main parameters used to find the optimal trajectory. A simulation framework to determine the optimal trajectory was developed in which the ego car travels in a highway environment scenario, using an agent-oriented approach. The performance of the algorithm was compared against optimal trajectories computed offline with the hybrid A* algorithm. The new framework provides trajectories close to the optimal trajectory and is computationally achievable. The agents were shown to follow safe and fast trajectories in three tests scenarios: emergency braking, overtaking and a complex situation with multiple vehicles around the ego vehicle. Different driver profiles were then tested in the complex scenario, showing that the proposed approach can adapt to driver preferences and provide a solution close to the optimal solution given the defined safety constraints.

Studies and field implementations demonstrate that variable speed limit (VSL) systems help freeway traffic congestion mitigations. This study presents a cycle-based variable speed limit (CVSL) strategy that could variate speed limits in the upstream segment of freeway merging areas. CVSL reduces speed limits in a fraction of a cycle. This variation creates artificial gaps on freeway mainlines, which in turn increases merging opportunities for incoming on-ramp vehicles. CVSL retains VSL's benefit in reducing delays while overcoming its shortcoming of no significant throughput increase. The authors establish analytical CVSL delay models, including freeway mainline and ramp delays with CVSL speed as a decision variable, and propose a robotic solution method to minimise the delays. The delay model and solutions are implemented and independently evaluated in microscopic traffic simulation. Particularly, the CVSL system is implemented and interfaced with the transportation flow open-source microscopic model (ETFOMM). Detector information from ETFOMM is fed into the delay model, and the optimised speed limit is fed back to ETFOMM. Two performance measures, total travel delay and throughput, are used to evaluate the CVSL system. Their simulation evaluation indicated up to 16% delay reduction and 6% throughput increase. This CVSL strategy is ready for field evaluation.

It is important for an urban rail transit system network to be resilient in the event of a disruption, and to be capable to recover in the shortest possible time. This study presents a holistic approach to analyse the survivability and recoverability of an urban rail–bus transit network during and after a disruption. A maximum survivability-minimum recovery time approach was formulated in this study to determine the number of affected passengers in the rail network during a disruption, the number of passengers who need to be transferred to alternate transport modes, and the recovery duration. A case study based on the Singapore urban mass rapid transit and bus networks is presented to demonstrate the applicability of the authors’ proposed framework. It was found from their analyses that the proposed framework could provide information on the state of rail network resilience given different disruption scenarios and estimate the recovery time after disruption occurrence.

A new vehicle positioning system is proposed using unscented Kalman filter for the data fusion of global positioning system and inertial navigation system, and a multi-hypothesis algorithm for map matching. The study presents a method to evaluate whether the results of the multi-hypothesis map matching algorithm can be used for feedback, and a strategy to increase the positioning accuracy based on this feedback. As the number of hypothesis nodes in the multi-hypothesis map matching algorithm grows exponentially with time, which costs lots of computation time and memory, several methods are proposed to reduce the number of hypotheses nodes by improving the generation method of hypothesis nodes, pruning the branches of multi-hypothesis tree, eliminating and merging the redundant nodes. Field test results indicate that the system can achieve much higher accuracy with the feedback from map matching, and can greatly save the computation time and memory.

The perception of risk, as the expectation of being involved in a traffic accident, is evaluated mainly in a subjective manner being perceptions as highly individual, and depending on experiences with accidents. However, both individual driver characteristics and driving behaviour entail certain perception errors in the risk level evaluation; as a consequence, drivers are very often not aware of the actual risk they are taking. For this aim, the paper presents a methodology for measuring the driver's perception error in the traffic accidents risk level evaluation based on the comparison between a measure of the risk perception subjectively obtained, and an objective measure obtained from kinematic parameters defining the driving style. Starting from a procedure that describes the relationship between lateral and longitudinal accelerations and speeds, we classified car drivers’ behaviour as safe or unsafe. Then, we defined three levels of risk of being involved in a road accident (low, medium, and high risk). The subjective measure of the risk perception is obtained by the judgements of drivers regarding their driving behaviour.

Several works have proposed longitudinal control strategies enabling a vehicle to operate adaptive cruise control and collision avoidance functions. However, no integration with lateral control has been proposed in the current state of the art, which motivates the developments of this work. This study presents an integrated control strategy for adaptive cruise control with auto-steering for highway driving. An appropriate logic-based control strategy is used to create synergies and safe interaction between longitudinal and lateral controllers to obtain both lateral stability and advanced adaptive cruise control functionalities. In particular, an index is proposed to evaluate lateral motion of the vehicle based on previously published experimental studies on human driving. In order to handle unstable lateral motion of the vehicle, the desired acceleration is determined based on physical limitation in braking with cornering situations. Simulation results show that the proposed integrated controller satisfies the performance in terms of autonomous driving, path tracking and collision avoidance for various driving situations.

Accurate collision prediction algorithms are important to provide drivers reliable warning messages. This study introduces a novel approach for collision prediction at intersections. The algorithm involves the use of an index called ‘minimal future distance (MFD)’, which is defined to be a future distance (FD) between the subject vehicle and the primary other vehicle, and a two-level dynamic threshold for performing the collision prediction task. Real-time vehicle motion information and surrounding road geometry are utilised to forecast FDs and identify MFD within an upcoming time horizon. The dynamic threshold in both emergency warning and normal warning situations consists of two parts, a vehicle heading direction-related part and a speed value-related part. Potential collisions are determined by the comparison of MFD with different dynamic thresholds in different driving scenarios. The combined use of vehicle real-time state and road geometry in the algorithm significantly increased the collision prediction accuracy. Furthermore, the use of dynamic thresholds ensured the promptness and robustness of the collision warning system. Simulation results show that the false positive rate and false negative rate of severe collisions at intersections can be robustly eliminated and those of marginal collisions can be kept low at 2.4 and 3.6%, respectively.

In this study, the authors improve the faster criterion in vehicle routing by extending the bi-delta distribution to the bi-normal distribution, which is a reasonable assumption for travel time on each road link. Based on this assumption, theoretical models are built for an arbitrary path and subsequently adopted to evaluate two candidate paths through probabilistic comparison. Experimental results demonstrate the bi-normal behaviour of link travel time in practice, and verify the faster criterion's superiority in determining the optimal path either on an artificial network with bi-normal distribution modelling link travel time or on a real road network with real traffic data. This study also validates that when the link number of one path is large, the probability density function of the whole path can be simplified by a normal distribution which approximates the sum of bi-normal distributions for each link.

In this study, the problem of electric multiple unit (EMU) circulation planning and train timetabling is studied through a coordinated process that provides a feedback mechanism to simultaneously minimise the number of EMUs, the number of EMU maintenance tasks and the train travel time. Based on an adjustable train departure time window as the key parameter in the problem, the approach mainly consists of two components: a column generation process, which searches for better EMU circulation routes; and a mathematical model, which produces timetables. These components interact according to the results of computational assessments until the solution reaches a certain level of optimality or the allotted computation time is exhausted. Finally, the authors test this approach using a small example to illustrate the effectiveness, and they also study a real-world case. A quantitative comparative analysis shows that long travel distances and travel times of trains significantly affect the number of EMUs used. The results indicate that the proposed model and algorithm can effectively address the coordinated optimisation problem of integrating EMU circulation planning and timetabling.