Abstract
Recently, the recognition of human activities has become a task of high interest within the wireless network community. Some Human Activity Recognition (HAR) systems have been developed for medical, military and security field, such as CARM and WiTrack. The importance of summarizing these existing works is to help researchers quickly familiarize themselves with this emerging field. A description of the main components of any HAR system is presented at first. Then we discuss general architecture of existing RF-signal based HAR system. We propose a twolevel taxonomy in accordance to RF-signal based HAR systems. We then discuss techniques and algorithms used by state-of-the-art HAR systems under each category and evaluate them in terms of recognition performance, recognition range, activity detected, and data source. Finally, we give a conclusion and present some open problems and ideas that, should be addressed in future research.
Original language | English |
---|---|
Title of host publication | Mobile Networks |
Subtitle of host publication | Concepts, Applications and Performance Analysis |
Publisher | Nova Science Publishers, Inc. |
Pages | 109-137 |
Number of pages | 29 |
ISBN (Electronic) | 9781536121551 |
ISBN (Print) | 9781536121230 |
State | Published - 1 Jan 2017 |
Keywords
- Human recognition systems, machine learning
- Radio-based sensing
- Signal processing