Section: New Results

Situation Aware Services on Mobile Devices

Participants : James Crowley, Thibaud Flury.

Modern mobile devices, such as smart phones and tablets, combine a rich set of sensors, internet connectivity, with embedded computational power and memory. The PRIMA group has recently demonstrated that it is possible to construct embedded software that uses the full suite of mobile sensors to recognise activities and learn the daily routines of users.

A first proof of concept has recently been constructed using recognition of places and activities. The system was trained by having student volunteers carry a cell phone running a data acquisition program that recorded signals from accelerometer, gyroscope, ambient sound, ambient light, Cell tower, wifi, bluetooth, and GPS based geolocalisation. The data were labeled by the students with ground truth data about transportation modes, places, and activities. This data was then used to learn recognition routines. Recognition of places, activities, and transportation was used to construct probabilistic models of daily routines using PRIMA's situation modelling techniques, previous demonstrated in constructing situation aware services. The system was demonstrated by constructing a Twitter Bot (a robot that publishes on twitter) that published information about volunteers during their daily activity.

A professional quality software system named CAM - Context Aware Manager - is currently under construction, and will be licensed to the PRIMA startup SItu8ed, for use in context aware mobile services.