UTS Data Science Institute has significant expertise in integrated predictive analytics to better manage large transportation networks, traffic congestions and infrastructure assets.

The team has a strong focus on developing advanced data-driven models for addressing future areas of smart city liveability, and the transition towards new transportation modes.  

Our transport and traffic capabilities include:

  • predicting traffic delays and propagation of delays
  • automated incident detection
  • automated inference of public transport demand from CCTV footage
  • structural monitoring of transport infrastructure
  • modelling of vehicle parking behaviour
  • micro-simulation of traffic movement
  • data fusion and real-time data visualisation.

The performance of contemporary public transport systems depends on profoundly complex interactions across passenger demand, network design, maintenance scheduling, incidents, time-tabling, vehicle availability and more.  We have developed software solutions that fuse these types of data streams and then leverage machine learning to deliver concrete and clear insight into the key predictors and drivers of public transport performance.

We have worked closely with Sydney Trains to develop new machine-learning-based vision processing systems that translate existing CCTV footage feeds into data streams that describe passenger flow throughout the terminal, counts of passengers at each platform, and the number of passengers boarding and exiting trains.  The system has already been trialed with a specific inner-city section of the Sydney rail network and will ultimately facilitate real-time customer impact analysis, dynamic delay propagation estimation, and advanced (dynamic) time-tabling and scheduling.

We have also used real-time GPS data, bus route information, and historical delay data to build real-time delay estimates for, and visualisations of, Sydney’s bus network. And with V/Line, a massive historical dataset capturing performance information from across their rural rail network was used to identify root causes of delays, provide quantified impact assessments, and to highlight potential emergent delay risks.