Skip to content Skip to navigation

Deep Context Sensing

Smartphone devices are getting increasing popular and users are also spending more time on them.


Goals: Developed a platform to collect useful and accurate user behaviour and actions in a manner that:
  • Is battery resource efficient.
  • Is network resource and cost efficient.
  • Does not impact the user experience on their devices.
Research Directions
Efficient Sensing
Collection of large diverse sets of data from different sources at minimum cost to battery. Leveraging:
  • Effective use of sensors
  • Collaboration, Participation & Cloud
  • Users behaviour
Efficient Network Usage
Transmission of valuable and private data at reduction of cost (network & charges). Leveraging:
  • Data type & behaviour
  • Exploiting context
  • Network protocols
Learning the User
Exploiting user patterns can further achieve reduction in savings in sensing as well as network usage.


The Collection Platform

Deep Context Collector
  • Mobile Agent for Gathering Pervasive Information & Evaluator/Experiment System (MAGPIES)
  • Android & iOS platform
  • Two operation modes:
Precision Mode: High-Medium fidelity data


Coarse Mode: Low-Medium fidelity data


Dynamic Collector Control
  • Control / Information / Notifications messages for
    real-time on-the-fly control to alter efficiency and
    fidelity of collector.
Interesting Apps
  • Novel and interesting mobile Apps can be build on top of the MAGPIES platform.
  • Partner apps that make uses of inferred context.

Last updated on 03 Sep 2015 .