| – |
– |
Development of dynamic driver behaviour models. The simplicity and accuracy of the agent-based algorithms, compared to conventional driver behaviour models, provides a means for determining the critical factors that influence drivers’ behaviour and modelling their decisions under the influence of traffic information |
| – |
|
Microscopic traffic simulation and the development of agent-based car following models. This research demonstrated for the first time the feasibility of developing Neugent (neural agent) car following algorithms for modelling the longitudinal interaction of vehicles in a traffic simulator. The Neugent model results showed better replication of car following behaviour than conventional models |
| – |
|
Development of adaptive agent-based algorithms for traffic signal control. In this research, multi-agent algorithms are used to coordinate signalised intersections. Each single intersection is modelled as a local autonomous controller equipped with local optimisation model. The results showed better performance compared to fixed time control and provided better throughput across the intersections |
| – |
|
Automated Incident Detection. This research resulted in the development of neural network algorithms for automatic detection of incidents on motorways. The algorithms are currently running online in a number of Transport Management Centres in Australia. In addition to their safety benefits, these systems were found to improve motorway efficiency by reducing the time to detect incidents by around 40 per cent |
|
|
Blank Line |
