A system for user task monitoring and assistance in ambient intelligent settings

Abstract

Existing task models are generally static (not used at runtime) and are used for the design or predictive evaluation of interactive systems. We propose to use the task model at runtime, in order to monitor user actions, check that they have not made any mistakes and give help when needed. We present a task model suitable for ambient environments that dynamically assigns states to tasks at the runtime. We also describe a monitoring and assistance system that uses our dynamic task model. Finally, we present a validation of our system through a simulation that shows how the interactions with the task model at runtime results in a dynamic system capable of providing assistance to users while they are carrying out their daily tasks.