Towards a knowledge base for activity recognition of diverse users

Abstract

This research work discusses a mathematical foundation based on probability theory and related disciplines for development of a knowledge base that would list the exhaustive ways or approaches, arising from universal diversity, by which any activity can be performed by a given user. The global challenge in this field is to address the needs associated with the constantly increasing population of the world who are diverse in multiple ways, through development of systems and frameworks that can serve as a long-term, robust, feasible, easily implementable, sustainable and economic solution for making the future of technology laden living environments ‘aware’ of the diverse ways by which users could be interacting with the environment and its components, in the context of performing Activities of Daily Living (ADLs), which are essential for their sustenance. Addressing this challenge serves as the main motivation for this work. Several case studies on different ADLs by application of the proposed framework were performed for development of the knowledge base and one study is presented here and discussed. In the context of making IoT-environments ‘Activity Aware’ for diverse users, such a knowledge base is expected to serve as a foundation to provide necessary information for various critical applications, for instance (1) user-centered activity recommendations by activity recommendation systems, (2) personalized behavior interventions for users with various forms of impairments – physical, cognitive etc. and (3) increasing the performance accuracy of the existing works in this field for adapting and responding to diverse interaction patterns exhibited by different users.