Title: Deriving Bespoke Human Activity Recognition Systems for Smart Homes

Date: Friday, April 14, 2023

Time: 11:00 AM - 1:00 PM ET

Location (in-person):  CODA C1315

Location (remote):   click here to join via Zoom

Join our Cloud HD Video Meeting

Zoom is the leader in modern enterprise video communications, with an easy, reliable cloud platform for video and audio conferencing, chat, and webinars across mobile, desktop, and room systems. Zoom Rooms is the original software-based conference room solution used around the world in board, conference, huddle, and training rooms, as well as executive offices and classrooms. Founded in 2011, Zoom helps businesses and organizations bring their teams together in a frictionless environment to get more done. Zoom is a publicly traded company headquartered in San Jose, CA.

gatech.zoom.us

Shruthi Hiremath

Ph.D. Student, Computer Science

School of Interactive Computing, Georgia Tech

Committee:

Dr. Thomas Plötz (advisor), School of Interactive Computing, Georgia Institute of Technology 

Dr. Gregory Abowd, School of Interactive Computing, Georgia Institute of Technology & College of Engineering, Northeastern University

Dr. Sonia Chernova, School of Interactive Computing, Georgia Institute of Technology 

Dr. Diane Cook, School of Electrical Engineering and Computer Science, Washington State University

Dr. Uichin Lee, School of Computing, Korea Advanced  Institute of Science and Technology

Abstract:

Smart Homes have come a long way: From research laboratories in the early days, through (almost) neglect, to their recent revival in real-world environments enabled by the existence of commodity devices and robust, standardized software frameworks. With such availability, human activity recognition (HAR) in smart homes has become attractive for many real-world applications, especially in the domain of Ambient Assisted Living (AAL). Yet, building an activity recognition system for specific smart homes, which are specialized spaces, with varying home layouts and inhabited by individuals with idiosyncratic behaviors and habits, is a non-trivial endeavor. 

My thesis addresses these challenges by first defining the Lifespan of a HAR system for smart homes, comprising three phases: i) bootstrapping; ii) update and maintenance; and iii) routine assessment. The Lifespan comprises components that are used to derive functional HAR systems quickly with minimal yet targeted involvement of residents. This is achieved through building novel analysis procedures and data analysis pipelines corresponding to the aforementioned phases. 

The contributions of my thesis are three-fold. First, I develop an initial bootstrapping procedure aimed at addressing the beginning of the life span of HAR resulting in a system that is capable of recognizing relevant and prominent activities. Second, I build on the bootstrapped system and introduce an effective update and extension procedure for continuous improvements of HAR systems with the aim of keeping up with changing data patterns in the home. Finally, I propose to extend the system to assess activity routines in homes using the developed activity recognition system.