SpamJam: Adaptive Travel Alerts using continuous position monitoring

Real-time travel alerts are one of the most compelling ways of informing people how to make the best travel choices, and can also be used to encourage modal or route shifts to reduce congestion. Providing highly specific, relevant, personally tailored information at the point of decision empowers the traveller to make the best assessment of their current options.
However, to be effective, travel alerts have to be almost effortless to use. Current systems typically require users to configure their travel patterns in advance, requiring at least some investment of time. Furthermore, in real life there is often considerable complexity and variability in behavioural patterns of personal travel, which is not captured by one simple configuration.
The goal of this project is to make travel alerts both self-configuring and richer, based on automated observation of the actual behaviours of each traveller. By continuously tracking travellers' location, and using other data sources that reveal location or intent, such as journey planning queries or calendar entries, a personal travel map can be built up, with regular journeys and patterns of variation inferred by the system. This data can then be used to dynamically configure subscriptions and switch them on and off to match the actual behaviour of the user on a particular day.
SpamJam is innovative in bringing together advances in the capabilities of mobile devices with the latest generation of alert engines to create a learning, adaptive system that can dynamically manage travel alerts.
Project partners are Kizoom, O2, Imperial College, Dynamical Systems Research and Tactical Systems Designers.