An Improved Approach for Complex Activity Recognition in Smart Homes

Abstract

The essence of intelligent assistive technologies in smart homes can be outlined as their ability to enhance the user experience in many ways. A way to accomplish this goal is to make such systems aware and knowledgeable about user interactions in the context of the given environment. In the context of smart homes, the ability of such intelligent systems to understand and analyze human behavior in the context of Activities of Daily Living is very essential. Modern day software-based solutions for development of such intelligent systems, should not only be able to sense human movements through Internet of Things based technologies but they should also be able to analyze these components of user interactions in a manner that facilitates optimal utilization of system resources, increases software productivity, improves system interoperability, reduces development and maintenance costs and allows both internal and external reusability of the solutions. Therefore, this paper discusses a study on the ‘Complex Activity Recognition Algorithm’ [1] and proposes a systematic approach to improve this algorithm for better performance and system optimization. The improved version of this algorithm has been tested on several activities and the results obtained uphold the relevance for implementation of the same in different real-time environments.