Autonomous systems that recognise, explain, and predict complex human activities
This project is part of a larger research initiative at Umeå University of 8 postdoc projects on autonomous systems for the future of industry and society.
The project is focused on the development of knowledge representation and automated reasoning techniques for recognising, explaining, and predicting complex human activities. Knowledge representation and automated reasoning techniques for performing this include temporal reasoning about goals, nonmonotonic causal logics, answer set programming, the treatment of incomplete, uncertain and inconsistent knowledge.
Understanding human activities and the context in which they take place are challenging, since a person's activity is driven by goals, motives, needs and norms that may be conflicting in a situation. Moreover, activities may be overlapping in time and performed differently depending on constraints given by a context. Therefore, sound computational theories for dealing with complex human activities are needed in order to develop a new generation of autonomous ambient intelligence systems.
Such context-awareness that is centered on the human and their activities enables the development of personalized and autonomously adaptive intelligent environments. The aim is to facilitate enhanced activity performance, for instance, in older adults who strive to continue living independently in their home environments.