AI in transportation modeling
Rick Williams (Movement and Place Consulting)
Session: 5
Room: 3010
Format: Presentation
Summary: How can AI contribute to transportation? Utilising an existing model to simulate the outcomes of potential changes, estimation, and mode choice models are some ways AI can assist. An illustrative application of data utilization is evident in Victoria, where traffic data and Scats are employed to enhance system optimization. The data requirements depend on the specific objectives we aim to achieve. For example, determining whether travel velocity data in a network is necessary or not. Data gathering poses a crucial consideration, particularly with regard to privacy. In cases of individual public transport usage, people may have concerns, but this issue is less relevant in the context of traffic signal data as it constitutes public information.
Machine learning can improve data quality, and remove inconsistency in data and result.
AI can handle OD estimation and cope with not aggregated data.
AI would predict and categorize demand by socio-demographic information.
In international events, ML can classify international and Australian people.
When we want to work on data and AI we should be careful about privacy and sharing information. The NSW government should make their travel household survey datasets open access.