Georgia Institute of Technology
Robust Human Activity Cataloguing through MmWave Sensing
Abstract
dc:description.abstractInnovations with mmWave radar sensing have enabled applications in Human Activity Recognition in a privacy-friendly way. Yet, its deployment into practice is hindered both by the lack of efficacy in cataloguing exercises in scenarios when users do not face the radar and the lack of ability to accurately segment and count exercises. Bridging those gaps, however, must be done in a low overhead manner to maintain practicality in deployment. This work presents RF-HAC, one of the first systems to deploy an automated, real-time, and robust view-agnostic Human Activity Cataloguing system into practice. RF-HAC achieves this by seamlessly integrating a self-supervised exercise classification model that leverages the diversity of multiple radars with robust doppler-based segmentation algorithm in a synergistic manner. This thesis will highlight the results of each of those components as well as proving its real-time operation.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Electrical and Computer Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Liu, Alan
- Advisor dc:contributor.advisor
-
- Sundaresan, Karthikeyan
- Committee members dc:contributor.committeemember
-
- Dhekne, Ashutosh
- Sivakumar, Ragupathy
- Weitnauer, Mary A.
Subjects
dc:subject × 2Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/77736